First Quarter RecommendationsThe NSCAI’s first quarterly report of 2020 is a compendium of recommendations from acrossthe Commission’s lines of effort, which correspond to the Tabs attached to this memo andsummarized below. It is not a comprehensive follow-up to the interim report and does not coverall areas that will be included in the final report. Rather it focuses on some areas theCommissioners identified as being most in need of attention, ripe for action, or foundational—and, therefore, important to outline now. It spells out recommendations that can inform ongoingdeliberations tied to policy, budget, and legislative calendars.A dynamic technology and dynamic national security environment require a dynamic approach tothe Commission’s work. Holding thoughts and ideas in abeyance for the next year until theCommission’s final report will not serve the national interest or enhance our security. The NSCAIis on track to submit its final report in March 2021. However, the pace of AI development, thegeopolitical situation, and the relevant authorization and budget timelines in 2020 representimportant opportunities for the Commission to contribute to ongoing efforts to foster researchand development, accelerate AI applications, and responsibly grapple with the implications of AIfor our security, economy, and society. We have concluded that as we become confident inspecific judgments and recommendations, the Commission should introduce them to the publicand stakeholders on a quarterly basis.National Security Commission on Artificial IntelligenceNSCAI_86a0fcc2-86a6-11ea-9377-66641c83ea00NSCAI CommissionersDR. ERIC SCHMIDTChairmanHON ROBERT O. WORKVice ChairmanSAFRA CATZDR. STEVE CHIENHON MIGNON CLYBURNCHRISTOPHER DARBYDR. KENNETH FORDDR. JOSÉ-MARIE GRIFFITHSDR. ERIC HORVITZANDREW JASSYGILMAN LOUIEDR. WILLIAM MARKDR. JASON MATHENYHON KATHARINA MCFARLANDDR. ANDREW MOOREThe U.S. wins the AI competition._86a10b18-86a6-11ea-9377-66641c83ea00To make recommendations to match the pace of change in AI_86a10ec4-86a6-11ea-9377-66641c83ea00TraceabilityAuditabilityAccountabilityR&DIncrease AI R&D Investments_86a11216-86a6-11ea-9377-66641c83ea001Academic CentersThe Commission assesses that the government should immediately double non-defense AI R&D funding to $2 billion for FY 2021 in order to strengthen AI at our academic centers and national laboratories, and advance development across the broad network of government-sponsored and affiliated laboratories around six key areas in the field where government support is critical to lay the foundation for our nation’s future security.National LaboratoriesRecommendations to Increase AI R&D Investments -- In Q1, the NSCAI focused on the FY21 R&D budget and identified funding recommendations that could have the greatest immediate impact on the United States’ ability to maintain global leadership in AI R&D. Federal funding for AI has not kept pace with the growth or the potential of the field...This will support efforts to apply AI in government, develop technical and ethical standards for AI technologies, expand AI across fields of science and medicine, and build vital public-private partnerships. The government should also launch a task force study and pilot program to establish a National AI Research Resource in order to democratize AI by ensuring wider access to resource-intensive computation and large, curated data sets.Non-Defense FundingDouble non-Defense AI R&D Funding for FY 2021_86a1173e-86a6-11ea-9377-66641c83ea001.1CongressProposed Legislative Action -- The goal of this initial boost in funding is to meet the demonstrated need for resources across the AI R&D community in a manner that does not flood the environment, but leverages a variety of existing vehicles and pathways to strengthen basic and applied research, build talent, and bolster existing national research assets to embrace and apply AI. The Commission sees this as the first of several, successive increases to gradually build the nation’s capabilities in AI research.Given our understanding of the needs of the R&D environment and the strengths of various agencies, the Commission recommends investments in NSF, DOE, the National Institute for Standards and Technology (NIST), NIH, and the National Aeronautics and Space Administration (NASA); and an expansion of a range of existing fellowship vehicles to support students and faculty pursuing AI-related degrees and research. This approach provides a stimulus to strengthenAI in academia, deepen work at our national laboratories, and advance development across the broad network of government-sponsored and affiliated laboratories. It will support efforts to apply AI to federal agency missions, develop technical and ethical standards for AI technologies, expand AI across fields of science and medicine, build public-private partnerships, anddemonstrate the positive potential of AI to the American people.The Commission recommends that Congress increase AI funding by approximately $1 billion18 and allocate the funding as follows:* $450 million to NSF,19 which is the preeminent science funding agency best poised to support the academic research community. These funds would support foundational AI research, research into semiconductors and microelectronics necessary for AI-related hardware,20 and expand NSF’s National AI Research Institutes Program and Cloudbank Program.21* $300 million to DOE,22 which oversees the nation’s 17 national labs where a wide range of basic and applied research is conducted—in both open and classified environments.These national assets also host specialized high-performance computing andexperimental facilities that can serve to accelerate the development of not only AI as a technology, but also novel hardware to support the evolution of AI and new AI-based applications critical to national mission areas. These labs also serve as valuable vehicles for public-private partnerships.* $125 million to NIH, which has a proven track record of funding researchtransformational to U.S. public health and society. Opportunities include harnessing AI methods to address public health emergencies, such as the rise of new pathogens and pandemics, including applications in precision medicine for critical care, advanced therapeutics, logistics, and epidemiology. Additional funding for NIH efforts would expand efforts to apply AI to biomedicine and to neuroscience, which promise to have synergies with next generations of AI and AI-specific hardware, among other opportunities.23* $50 million to NIST, which plays a critical role in developing standards with U.S. industry. These funds would enable NIST to lead the setting of protocols and nation-wide testing standards for AI, alleviating the cost burden from compliance with multiple standards at the local, state, and federal levels, and accelerating development and fielding of the technology. NIST would be able to expand its efforts to advance the tools to measure and understand the capabilities and limitations of AI technologies and its underlying data,building the confidence and trust needed to expedite technology adoption, deployment, and use.* $75 million to NASA, to leverage its existing expertise in operational deployment of AI in space missions. Due to the urgent need for autonomous operations and AI in space domain24 and the criticality of space to national security, we recommend this additional funding to enable NASA to expand its development and application of AI techniques tailored to the space domain and continue to serve as a pathfinder providing operational experience with AI. This work has direct applications to national security.25* $100 million to expand select fellowship and scholarship programs managed by DoD, DOE, NASA, and NSF.26 Additional funding through these vehicles would support additional undergraduate and graduate students to pursue AI-related fields of study, helping to strengthen academia, grow the domestic talent pipeline, and provide pathways into government for technical talent. Similarly, career/faculty fellowship vehicles supporting researchers in academia can serve to stem the flow of researchers to industry and invest in top talent to pursue big ideas. ^The Commission recommends that Congress mandate reporting on the additional AI scholarships and fellowships awarded as a result of these increases.AI R&D CommunityNSFDOENational Institute for Standards and Technology (NIST)NIHNational Aeronautics and Space Administration (NASA)The Commission recommends that Congress roughly double the funding level of non-defense AI R&D for FY 2021 to begin to immediately address the funding deficit in the field and build the capacity for compounding higher levels of funding and investment in future years. This funding should increase agency topline levels, not repurpose funds from within existing agency budgets, and be used by agencies to fund new research and initiatives, not to support re-labeled existing efforts. We also recommend investments in specific fellowships administered by DoD and the Defense Advanced Research Projects Agency (DARPA), as well as in the national labs managed by DOE’s National Nuclear Security Administration, which falls under the National Defense Budget. In Tab 4, we have also outlined recommended investments in DARPA’s microelectronics program and recommend the establishment of a new microelectronics program in the Intelligence Advanced Research Projects Activity (IARPA), both with the goal of securing U.S. advantages in AI hardware.17PrioritiesPrioritize Funding for Specific Areas of AI_86a119c8-86a6-11ea-9377-66641c83ea001.2The Commission recommends that additional defense and non-defense AI research funding should be applied to advance six key areas in the field. In these areas, the Commission assesses that government support is critical to lay a foundation for our nation’s future security.The Commission recommends funding be prioritized for the following areas ...This is by no means a comprehensive list, but rather an identification of near-term priorities in which government investment could have an outsized return on investment and advance the field toward future capabilities that will underwrite our national security and defense.DirectionDirect machine learning._86a11c98-86a6-11ea-9377-66641c83ea001.2-1Novel machine learning directions. To further non-traditional approaches to supervised machine learning in an unsupervised or semi-supervised manner as well as the transfer of learning from one task or domain to another.SupervisionAdvance machine learning in an unsupervised or semi-supervised manner._86a11fb8-86a6-11ea-9377-66641c83ea001.2-1aLearning TransferTransfer learning from one task or domain to another._86a1227e-86a6-11ea-9377-66641c83ea001.2-1bTEVVTest, evaluate, verify, and validate and build checks and balances into an AI system._86a1254e-86a6-11ea-9377-66641c83ea001.2-2Testing, Evaluation, Verification, and Validation (TEVV) of AI systems. Todevelop a better understanding of how to conduct TEVV and build checks and balances into an AI system.RobustnessCultivate more robust methods._86a12e9a-86a6-11ea-9377-66641c83ea001.2-3Robust machine learning. To cultivate more robust methods that can overcome adverse conditions, and advance approaches that enable assessment of types and levels of vulnerability and immunity.AdversityOvercome adverse conditions._86a13c82-86a6-11ea-9377-66641c83ea001.2-3aVulnerability & ImmunityAssess types and levels of vulnerability and immunity._86a1410a-86a6-11ea-9377-66641c83ea001.2-3bInteractionUnderstand interacting AI systems._86a145a6-86a6-11ea-9377-66641c83ea001.2-4Complex multi-agent scenarios. To advance the understanding of interacting AI systems, including the application of game theory to varied and complex scenarios, including interactions between cohorts composed of a mixture of humans and AI technologies.Modeling, Simulation & Design_86a14f38-86a6-11ea-9377-66641c83ea001.2-5AI for modeling, simulation, and design. To progress the use of rich simulations as a source of data and scenarios for training and testing AI systems, and to use AI to solve complex analytical problems and to serve as a generative design engine in scientific discovery and engineering.Training & TestingUse simulations for training and testing AI systems._86a15384-86a6-11ea-9377-66641c83ea001.2-5aAnalytical ProblemsSolve complex analytical problems._86a1580c-86a6-11ea-9377-66641c83ea001.2-5bDiscovery & EngineeringDesign scientific discovery and engineering._86a16180-86a6-11ea-9377-66641c83ea001.2-5cEnvironments & ScenariosEnhance actionable awareness and insight across a range of complex, dynamic environments and scenarios._86a1663a-86a6-11ea-9377-66641c83ea001.2-6Advanced scene understanding. To evolve perceptual models to incorporate multi-source and multi-modal information to support enhanced actionable awareness and insight across a range of complex, dynamic environments and scenarios.Study & PilotLaunch a Task Force Study and Pilot Program to Establish a National AI Research Resource._86a16b12-86a6-11ea-9377-66641c83ea001.3CongressProposed Legislative Action -- Congress should authorize a task force, co-led by NSF and the Office of Science and TechnologyPolicy (OSTP), that includes representatives from key government agencies anticipated tocontribute data sets to the NAIRR platform. The task force—composed of government leadersand representatives from the academic and commercial AI communities—would produce aroadmap laying out ownership, governance, capabilities, and sustainment of the national resourcewithin 180 days of enactment. Once the roadmap is developed, funding for and implementationof a multi-year pilot would transition to the responsible agency determined through the study.The roadmap would define the following:* Ownership: Agency or organization responsible for the implementation, deployment, on-going development, and staff support for the NAIRR.* Governance: Composition of the structural processes and oversight body needed to establish strategic direction, make programmatic decisions, and manage the allocation of resources. For example, defining the grant-making selection and allocation processes, selecting government and non-government AI-ready datasets to add to the NAIRR over time, and determining permissions and management of access to the data and resources hosted on the platform.* Capabilities: Novel, scalable cloud architecture designed to enable secured access control; resident data engineering and curation expertise; provision of data sets that are findable, accessible, interoperable, and reusable; computation on hosted data that does not leave the platform; educational tools and services; and a user interface portal. The pilot implementation provides a proving ground for developing data interface and quality standards, curation best practices, anonymization techniques, and standardized procedures and criteria for determining which government and non-government data can be made publicly available under what conditions.* Sustainment: Business model based in public-private partnerships that can support provision of scalable cloud resources and a staff of cloud architects, data engineers, and educational/technical support, all at a minimal cost to the researcher. ^Funding for the pilot would support staffing of the program and the cloud resources (persistent data storage and compute services), augmented through public-private partnerships. Staff would be responsible for maintaining and improving the architecture solution, curating data sets, building interfaces and tools, and providing support to researchers as they use the NAIRR.NSFOffice of Science and Technology PolicyTo advance AI research in the United States, Congress should authorize and appropriate $25 million in funding for the first year of a five-year pilot program to develop, implement, and sustain a National AI Research Resource (NAIRR) infrastructure that would provide researchers and students with access to compute resources, co-located with AI-ready government data sets,educational tools, and user support.29 This infrastructure would leverage public-private partnerships and build on existing government efforts, avoiding high start-up costs of a government-run data center.The NAIRR pilot program would accelerate and strengthen AI research across the U.S. by removing the high-cost barrier to entry of compute resources while also fueling research and training through access to machine learning-ready data sets of real-world, representative U.S. government data.30 Combined with the provision of training tools and support options, this infrastructure would democratize AI R&D outside of elite universities and big technology companies and further enable the application of AI approaches across scientific fields anddisciplines, unlocking breakthroughs.31DoD ApplicationsAccelerate AI Application in DoD_86a1730a-86a6-11ea-9377-66641c83ea002DoDThe Commission recommends that DoD and the Office of the Director of National Intelligence (ODNI) establish a Steering Committee on Emerging Technology tri-chaired by the Deputy Secretary of Defense, the Vice Chairman of the Joint Chiefs of Staff, and the Principal Deputy Director of ODNI in order to drive action on emerging technologies that otherwise may not be prioritized.Deputy Secretary of DefenseVice Chairman, Joint Chiefs of StaffOffice of the Director of National IntelligencePrincipal Deputy Director, ODNISteering Committee on Emerging TechnologyThe Steering Committee would assess novel threats related to emerging technology; connect strategic vision to organizational change; focus concept and capability development on emerging threats; guide defense investments that ensure America’s strategic advantage against near-peer competitors; and provide authority to drive technology adoption and application.Director, Joint Artificial Intelligence CenterThe NSCAI also recommends that the Director of the Joint Artificial Intelligence Center (JAIC) should report directly to the Secretary of Defense to ensure senior leadership oversight of DoD AI efforts and alignment with Department priorities, and that the JAIC should continue to be led by a three-star general or flag officer with significant operational experience.Q1 recommendations for applying AI focus on key organizational changes that can bolster existing initiatives and accelerate DoD AI application in the near term. The recommendations prioritize creating top-down leadership mechanisms to overcome structural barriers and enable strategic change.ET Steering CommitteeEstablish a Steering Committee on Emerging Technology._86a177a6-86a6-11ea-9377-66641c83ea002.1Secretary of DefenseProposed Executive Branch Action -- The Secretary of Defense and Director of National Intelligence should issue a directive immediately establishing the senior oversight committee described above. The Steering Committee provides a forum for the Office of the Secretary of Defense, the Joint Staff, and the Office of the Director for National Intelligence senior leadership to focus on developing concepts and capabilities that include emerging and disruptive technologies and meet current and future operational challenges facing the nation.Director of National IntelligenceDeputy Secretary of DefenseVice Chairman, Joint Chiefs of StaffPrincipal Deputy Director, ODNITri-Chaired Steering CommitteeThe Tri-Chaired Steering Committee should:* Review credible assessments of emerging threats and identify adversary investments and advances in emerging technologies such as AI/ML, 5G, quantum computing, autonomy, robotics, bio-technology, nanotechnology, and other technologies as appropriate.* Identify, prioritize, and resource technically and operationally feasible technology solutions to the most pressing operational challenges facing DoD as identified in the National Defense Strategy.* Direct changes in existing programs as well as direct sufficient resources forexperimentation and prototyping of one-off, high-risk, high-reward projects.* Propose changes to existing policies and processes that are not optimized for today’snon-linear, risk-taking technology innovation, which requires agility, creativity, and speed.Steering Committee MembersSteering Committee members should include DoD and ODNI executives, including but not limited to the Chief Management Officer, appropriate Under Secretaries of Defense and Deputy Directors of National Intelligence, Office of the Secretary of Defense Office of Cost Assessment and Program Evaluation (CAPE), the Directors of the Joint Artificial Intelligence Center (JAIC) and the Augmenting Intelligence Using Machines Initiative (AIM), and the Vice Chiefs of the military services.Chief Management OfficerUnder Secretaries of DefenseDeputy Directors of National IntelligenceOffice of the Secretary of Defense Office of Cost Assessment and Program Evaluation (CAPE)Director, Joint Artificial Intelligence Center (JAIC)Director, Augmenting Intelligence Using Machines Initiative (AIM)Vice Chiefsof the military servicesUnder Secretary of Defense for Research and EngineeringThe Under Secretary of Defense for Research and Engineering should provide technical expertise to the Committee, as aligned with their existing Departmental responsibilities.Service AI LeadsService AI leads should participate when AI is the primary topic of discussion.DoD and the Office of the Director of National Intelligence (ODNI) should establish a Steering Committee on Emerging Technology tri-chaired by the Deputy Secretary of Defense, the Vice Chairman of the Joint Chiefs of Staff, and the Principal Deputy Director of ODNI. Establishing such a tri-chaired committee will integrate DoD and IC AI efforts and provide the top-down focus needed for DoD to overcome the bureaucratic challenges impeding AI application. Specifically, it will help enhance intelligence analysis related to emerging technology, connect strategic vision to organizational change; focus concept and capability development on emerging threats; guide defense investments that ensure America’s strategic advantage against near-peer competitors; and provide the authority to drive technology adoption and application by the Department. It will also make recommendations directly to the Secretary of Defense on Departmental innovation priorities.Reporting & DelegationThe Director of the Joint Artificial Intelligence Center (JAIC) should report directly to the Secretary of Defense, who may delegate this authority to the Deputy Secretary of Defense._86a17c42-86a6-11ea-9377-66641c83ea002.2Director, Joint Artificial Intelligence CenterSecretary of DefenseProposed Executive Branch Action -- The Secretary of Defense should immediately issue a memo designating the Director of the Joint Artificial Intelligence Center as a direct report or delegate such authority to the Deputy Secretary of Defense.Recommendations from the JAIC to the Secretary of Defense regarding new applications of, and modifications to, existing and near-term capabilities that provide an operational advantage to the Department should be reviewed by the new tri-chaired Steering Committee and, where appropriate, conveyed as investment recommendations to the Deputy’s Management Action Group.Deputy Secretary of DefenseCongressProposed Legislative Branch Action -- Congress should use the FY 2021 National Defense Authorization Act to establish the JAIC as a direct report to the Secretary of Defense, who may delegate this authority to the Deputy Secretary of Defense.Director, JAICThe Director of the JAIC should report directly to the Secretary of Defense or his Deputy regarding the JAIC’s central roles and responsibilities.The JAIC’s current organizational placement within the Office of the Secretary of Defense and its fiscal relationship to the Defense Information Systems Agency should remain unchanged.DISADoD CIOHowever, the JAIC and the CIO should fully separate in their roles and responsibilities. The JAIC should hold primary responsibility for AI applications, while the CIO should continue to lead DoD’s broader digital transformation efforts.Strategic Capabilities OfficeA direct reporting line to senior leadership has precedent within the Department. The Strategic Capabilities Office (SCO), which also seeks to further defense innovation, was established in 2012.Since then it has held various reporting structures, including reporting directly to the Deputy Secretary. The direct reporting line gave SCO the visibility, support, and resources to build their capacity and prove their value to the services and to Congress. After organizational changes in recent years, Congress re-elevated SCO to report directly to the Deputy Secretary of Defense in the FY 2020 National Defense Authorization Act (Pub. Law 116-92, Sec. 233).This change ensures senior leadership oversight of DoD AI efforts and alignment with Department priorities by elevating JAIC’s reporting authority, which currently runs through the Chief Information Officer (CIO), to the Secretary of Defense or his Deputy. This recommendation seeks to give the Secretary of Defense, or Deputy by delegation, the ability to exercise authority and direction over the JAIC...Direct attention from senior leadership can drive organizational focus and empower coordinating entities to gain service support and move initiatives forward. Establishing Secretary of Defense authority and direction over the Department’s leading AI effort provides the requisite level of senior oversight and support needed to preserve the Department’s initial AI projects, enable theirgrowth, and ensure the Department can develop the capabilities needed to successfully adopt AI applications.Operational ExperienceMaintain the Director of the JAIC as a three-star general or flag officer with proven operational experience._86a1848a-86a6-11ea-9377-66641c83ea002.3Secretary of DefenseProposed Executive Branch Action -- The Secretary of Defense should require that the Director of the JAIC remain a three-star general or flag officer with significant operational experience.CongressProposed Legislative Branch Action -- Congress should use the FY 2021 National Defense Authorization Act to designate that the Director of the JAIC will be a three-star general or flag officer by creating a three-star billet forthe role.Three-star leadership allows the JAIC to engage with the services at a senior rank and within their command structure. Operational experience enables the Director to understand how AI can serve operational requirements and better communicate with the services as to how AI meets capability needs. This has helped the JAIC build inroads with the services as a new organization in the Office of the Secretary of Defense. Maintaining a strong relationship with the services is necessary going forward to enable the transition of AI applications to service programs as they become available and ensure application development meets service needs. Retaining this credibility is important in the short-term as the JAIC’s programs and capabilities mature. However, the JAIC should retain the flexibility to consider changing to civilian leadership in the future and ensure strong AI technical expertise exists in supporting roles.WorkforceStrengthen the AI Workforce_86a18a48-86a6-11ea-9377-66641c83ea003ExpertsDevelopersCyber Excepted ServiceHR ProfessionalsHiring ManagersOrganizational LeadersQ1 recommendations target recruiting experts and developers, training end users, identifying internal talent, and educating those in support roles. They are premised on the belief that if the government cannot improve hiring practices, baseline knowledge, recruitment, and talent exchanges, we will struggle to accomplish any significant AI progress. The NSCAI proposes expanding the Cyber Excepted Service; establishing AI Literacy course for Human Resource (HR) professionals; rebalancing the relationship between HR professionals, hiring managers, and organizational leaders; creating exemptions fromOPM qualification policies; accelerating security clearance reviews for AIpractitioners; expanding unclassified workspaces; piloting portfolio-based ratherthan resume-based hiring; mandating AI training; incentivizing self-development AI courses; incentivizing programming proficiency; including computational thinking in the Armed Services Vocational Aptitude Battery, or ASVAB; hiring professors part-time in government research labs; utilizing the pathways internship program; expanding the Cyber Corps/Scholarship for Service program; and increasing fellowships and partnership with industry. If adopted, the recommendations would cumulatively change technologist hiring and talent management across the national security enterprise.Cyber ServiceExpand the Cyber Excepted Service_86a18fca-86a6-11ea-9377-66641c83ea003.1Executive BranchProposed Executive Branch Action -- Extend the Defense Civilian Intelligence Personnel System Interchange Agreement (DCIPS IA) to personnel hired with the CES.37 Current Title 5 employees that voluntarily transfer to CES will generally have Title 5 status. New hires often will not, damaging their ability to stay in government or move to other positions within the government that are outside the CES. This will impede the USG’s ability to attract and retain talent. Extending the DCIPS IA will allow personnel hired under the CES to transfer to the competitive service with the same status as Title 5 employees.Office of Personnel ManagementCongressProposed Legislative Branch Action -- The Congress should use the National Defense Authorization Act to expand the CES to include AI positions. An expanded CES should allow the DoD and military services to hire AI practitioners more quickly. Hiring success metrics Congress should track should include: 1) how often the CES is used, 2) how often the CES is used for AI-related positions compared to the competitive service, 3) the number of job offers made using both the CES and the competitive service and their relative success rate, and 4) how quickly the DoD and military services hire AIpractitioners compared to positions that do not have direct hiring authorities or excepted service.Existing authorities for other national security departments and agencies are adequate as written, but not as widely used. Their challenge generally lies in receiving approval to use hiring authorities outside the competitive service. Organizations applying for excepted service status and direct hiring authorities from the Office of Personnel Management (OPM) frequently face barriers thateither draw out the application process or prevent organizations from using processes outside of the competitive service.38 Congress should require OPM to monitor the application process to evaluate the rate and speed at which exemptions to the competitive service are granted to applying agencies. OPM’s reporting requirements to Congress should include: 1) the time required tocomplete the application process, 2) the percentage of applying agencies that complete the application process, 3) the percentage of applying organizations allowed to use the CES, and 4) the reason organizations are disapproved.The government should expand the CES to explicitly include AI positions, which should include positions that accomplish tasks described by Section 1051(f) of the John S. McCain National Defense Authorization Act of Fiscal Year 2019 (Pub. L. 115-232), which provides a working definition of AI for defense and national security purposes. Expanding the CES rather than creating a new hiring authority avoids contributing to the confusing proliferation of authoritiesand avoids delays from the time required to establish and begin using a new hiring authority. The government should also allow personnel hired under CES to be appointed to competitive service positions without competing as external, non-federal applicants.It should be noted that cyber and AI are different fields. Expanding the CES to a Cyber and AI Excepted Service would avoid increasing administrative burdens and adding another hiring authority. This should not be taken as an indication that AI and cyber are synonymous, as the education and skills for each field differ. Also, the National Security Commission on AI believes the United States Government and Department of Defense may benefit from developing a digitalcorps, and is considering the steps needed to create one. The above recommendation is consistent with that end, and can be easily incorporated into a digital corps if the decision to pursue one is approved.AI LiteracyIncrease Human Resource Team AI Literacy_86a1984e-86a6-11ea-9377-66641c83ea003.2Human Resource TeamDepartment of DefenseProposed Executive Branch Action -- The Department of Defense should establish short courses for HR professionals, hiring managers, and recruiters that are or will be responsible for hiring software developers, data scientists, or AI practitioners. The course should address applying for and the use of available hiring authorities, including but not limited to direct hiring authorities,42 the Cyber Excepted Service,43 special government employees, Intergovernmental Personnel Act, and highly qualified experts.44 The course should also address the use of ePortfolio reviews (see below) and provide a general familiarity with the software development, data science, and AI fields. This includes anoverview of software development and business processes, data management practices related to machine learning, an introduction to machine learning and deep learning, an introduction to AI, and AI workforce roles and archetypes.Software DevelopersData ScientistsAI PractitionersCongressProposed Legislative Branch Action -- The Armed Services committees should use the FY 2021 NDAA to require the Department of Defense to establish a human resources software development, data science, and AI short course within one year of the NDAA’s enactment. Twenty percent of the HR teams, hiring managers, and recruiters responsible for hiring software developers, data scientists, or AI practitioners should be required to pass the course by the end of the first year, with the minimum percentage certified increasing by ten percent each year until 80 percent of the applicable human resource workforce is certified. Equivalent practices are already the norm in some top-tier technologycompanies.45 Appropriators should set aside $2.5 million of DoD O&M funding for the creation and execution of a short course for HR professionals, hiring managers, and recruiters.46HR TeamsHiring ManagersRecruitersThe Department of Defense, Department of Homeland Security, Federal Bureau of Investigation, and Intelligence Community (IC) should establish a training and certification program for HR professionals that ensures:* familiarity with their organization’s goals regarding AI;* hiring practices outside of the competitive service; and* software development, AI, and AI workforce literacy for HR teams, hiring managers, and recruiters. ^Agencies should audit short courses annually to ensure nomenclature, definitions, organization goals and vision, and other information maintains the accuracy and pace consistent with the evolution of AI.Hiring TriangleRe-balance the Hiring Triangle_86a19cea-86a6-11ea-9377-66641c83ea003.3HR DepartmentsAI PractitionersSubject Matter ExpertsSecretary of DefenseProposed Executive Branch Action -- The Secretary of Defense should issue a memorandum of intent that rebalances the hiring triangle.The memorandum should include the following:* A description of the flaws in the current hiring process for software developers, data scientists, and AI practitioners (see above);* A mandate for commanders to assume greater responsibility for the results of civilian hiring of software developers, data scientists, and AI practitioners;* Allow hiring managers to dictate the use of applicable direct hiring authorities,qualification standards, and to determine who is qualified;* An increase in the role of subject matter experts such as software engineers in the hiring process;* A reduction of the ability of HR professionals to determine if a potential employee is well qualified for software development, data science, or AI positions;* A desire to use direct hiring authorities and the Cyber Excepted Service when they are the fastest way to hire qualified personnel; and,* A requirement to report the status of software engineer and AI practitioner hiring in the organization, including hiring time, use of hiring authorities besides the competitive service, and the role of subject matter experts in hiring decisions.CongressProposed Legislative Branch Action -- The Armed Services committees should use the FY 2021 NDAA to require the Department of Defense to provide referral bonuses to software development, data science, and AI experts. DoD Instruction 1400.24-V451 already authorizes the use of referral bonuses for recruitment andhiring, but it is unclear how often it is used and to what effect. Performance metrics should include the number of times performance bonuses are awarded per year, both in gross numbers and as a percentage of hiring decisions for positions involving software development, data science, and AI. Performance metrics should also include hiring time and the use of hiring authoritiesbesides the competitive service.Appropriators should set aside $100,000 of O&M funding for each military service and for the Office of the Secretary of Defense to use as referral bonuses for software development, data science, and AI experts.HR departments should not be in charge of selecting AI practitioners for the government workforce. Instead, qualified practitioners within the government workforce should be responsible for selecting their own team members, HR teams should be responsible for executing hiring processes with little discretion about hiring choices, and commanders and civilian leaders should aggressively assume more responsibility for the output of the entire process.47 In addition,to reduce the burden placed on HR teams, subject matter experts should receive referral bonuses for civilian experts hired after applying based on the subject matter experts referral.Billets & PDsGrant Exemption from OPM General Schedule Qualification Policies for Specific Billets and Position Descriptions_86a1a1fe-86a6-11ea-9377-66641c83ea003.4OPMExecutive BranchProposed Executive Branch Action -- Two-star and above commands and their civilian equivalents should be authorized to declare individual billets and position descriptions exempt from OPM qualification standards without approval from OPM or any more senior authority within the DoD. Separately, OPM should create and execute a process by which federal departments and agencies can apply for billets or position descriptions to be exempt from general schedule qualification policies. Once the process has been created, it should be integrated into Human Resource Team AI Literacy training (see above).OPMSubject Matter ExpertsExpert TechnologistsCongressProposed Legislative Branch Action -- Authorizing committees should direct the Office of Personnel Management to amend 5 CFR §338.301, on service appointments, to allow service secretaries and cabinet officials to createexceptions from the Qualification Standards for General Schedule Positions by individual billet or position description.Even when subject matter experts are able to choose their applicants, they are constrained by OPM minimum qualification standards. While these are important and have increased fairness in hiring, they also prevent expert technologists that do not have master’s degrees, and in some cases bachelor’s degrees or comparable work experience, from joining the government at a reasonablelevel of compensation. For example, a 19 year-old software developer or AI practitioner might have a proven track record on cybersecurity or in AI competitions, but can only enter the government as a GS-7. To reduce this hiring challenge, the government should allow agencies to exempt certain billets from OPM general schedule qualification policies, and instead allow local hiring managers to make an independent decision about both hiring and the pay grade.Security ClearancesAccelerate Security Clearance Investigation and Adjudication_86a1abea-86a6-11ea-9377-66641c83ea003.5Defense Counterintelligence and Security AgencyProposed Executive Branch Action -- Upon request by a two-star and above commands or their civilian equivalents, the Defense Counterintelligence and Security Agency should accelerate the security clearance process for personnel being hired under the Cyber Excepted Service and for AI, data science, and software development positions, and set a standard that they have an interim top secret clearance within 30 days and an interim secret clearance within 20 days of receipt of the SF86. The process can be established with a separate, parallel infrastructure to other security clearances, or by prioritizing personnel hired under the Cyber Excepted Service.CongressProposed Legislative Branch Action -- The Armed Services committees should use the FY 2021 NDAA to require the DCSA to establish an accelerated security clearance timeline for personnel being hired under the Cyber Excepted Service and for AI, data science, and software development positions as defined in the John S. McCain National Defense Authorization Act of Fiscal Year 2019 (P.L. 115-232). The time required for these personnel to receive interim secret and top secret clearances should be included as a quarterly briefing or reporting requirement to the Armed Services committees.The USG should allow leaders of two-star and above commands and their civilian equivalents to prioritize personnel hired under the CES and for AI, data science, and software development positions during the security clearance process, and set a standard that they have an interim secret clearance within 20 days and an interim top secret clearance within 30 days. Prioritizing CyberExcepted Service hires will be contentious, as almost every organization believes its subject matter experts or leaders should be a high priority. There is a strong case that AI genuinely and uniquely requires prioritized security clearances. First, due to the small number of AI practitioners both in the United States and around the world and the degree of skill it requires, AI is a uniquely talent-dependent field. Second, the commercial and academic demand for AI practitioners is extremely high and competitive. AI practitioners that put their work on hold for months at a time to wait for a security clearance can, and sometimes do, simply turn to the private sector and are quickly hired at a salary the government cannot match. Third, senior leaders, including the Secretary ofDefense51 and the Principal Deputy Director of National Intelligence,52 have made AI their leading technology priority, but their departments and agencies have been unable to hire the talent they need to implement AI. These reasons support the argument that AI-related positions are genuinely a higher priority for security clearances than other positions.WorkspacesCreate Unclassified Workspaces_86a1b8f6-86a6-11ea-9377-66641c83ea003.6Executive BranchProposed Executive Branch Action -- The Secretary of Defense, Secretary of Homeland Security, and Director of National Intelligence should issue guidance directing their respective organizations to establish, within 180 days of publication, unclassified work spaces for AI practitioners at major commands, combatant commands, and major research facilities within 180 days of publication.Secretary of DefenseSecretary of Homeland SecurityDirector of National IntelligenceCongressProposed Legislative Branch Action -- The Armed Services committees should use the FY 2021 NDAA to require the DoD, Department of Homeland Security (DHS), and IC to create unclassified workspaces for uncleared AI practitioners at major commands, combatant commands, and major research facilities.DoDDHSAI PractitionersMajor CommandsCombatant CommandsMajor Research FacilitiesNational security departments and agencies need to establish facilities where employees waiting on security clearances can perform unclassified work until they receive a clearance. This will allow organizations to quickly hire new employees, even when their security clearance process takes longer.ePortfoliosUse ePortfolio Reviews_86a1c170-86a6-11ea-9377-66641c83ea003.7Executive BranchProposed Executive Branch Action -- The Department of Defense, Department of Homeland Security, and Intelligence Community should develop and implement an ePortfolio review process for software development, data science, and AI positions and invite applicants to submit ePortfolios.Department of DefenseTo implement this change, the Department of Defense should pilot the use of ePortfolio reviews in a major office or command.Department of Homeland SecurityIntelligence CommunityePortfolio Review TeamsAll ePortfolio review teams must have subject matter experts with skill sets related to the open position leading the review process.Subject Matter ExpertsIf SMEs do not assume a larger role in the hiring process and HR teams and hiring managers are not trained to read an ePortfolio, the process will be ineffective and potentially harmful.HR ProfessionalsHR professionals who are involved in ePortfolio reviews should receive training on both topics.Hiring ManagersCongressProposed Legislative Branch Action -- The Armed Services committees should use the FY 2021 NDAA to mandate the piloting of ePortfolio reviews in at least one major office or command. The pilot should be evaluated based on successful and unsuccessful applicant’s perception of the application process and existing employee’s perception of the effectiveness of the piloted ePortfolio review process. If the pilot has positive results, Congress should direct national security departments and agencies to review and update policies that inhibit the broad application of ePortfolio reviews rather than resumes during the hiring of technologists within one year of the completion of the pilot.When hiring AI practitioners, the Department of Defense, Department of Homeland Security, and Intelligence Community should use resumes for their initial screening process, then during the remainder of the evaluation process replace resume reviews with an ePortfolio review designed to evaluate data science, software development, and AI competency. This approach would be based on private sector best practices both at major software companies and at small start-ups.56 ePortfolio reviews show examples of an applicant’s previous development work as a tangible demonstration of his or her capabilities.57 To minimize applicant requirements and screener workload, an ePortfolio should contain three examples of the applicant’s best work. This recommendation is closely linked to the above recommendations that address the role of HRteams.TrainingRequire mandatory training to improve baseline AI literacy._86a1cf44-86a6-11ea-9377-66641c83ea003.8DoDDHSExecutive BranchProposed Executive Branch Action -- DoD and DHS should evaluate current commercial offerings for AI courses and determine the best fit based on their organizations’ needs. The course should be structured with successive levels of comprehension, annual repetitions, and certification for different levels of competency. The course should include baseline instruction in the nature, development, limitations, and application of AI and data science and the basics of data management. Instruction related to the responsible and ethical development and fielding of AI should also be included (see Tab 6, Recommendation 1). Both departments should also establish and implement a system for tracking course completion and compensating those who complete the course.DoDDoD and DHS currently struggle to implement AI solutions due to a lack of baseline knowledge, which, as described above, is a greater challenge to readiness than the requirement to conduct an annual training course.DHSU.S. ArmyThere is well-justified resistance to increasing mandatory training. The Army recently reduced its mandatory training requirements based on the belief that they distracted from combat-focused training, damaging readiness.60 While that is accurate for many mandatory training requirements in the past, it will not be the case for well-executed AI training. AI, if implemented successfully,will help organizations become more efficient and effective, and create new capabilities.CongressProposed Legislative Branch Action -- Authorizing committees should mandate that the DoD and DHS develop and implement mandatory annual training to improve AI literacy. For mandatory annual training, authorizing language should mandate course implementation, annual course assessment and revision, the development of course completion tracking, and the establishment of compensatory time-off for course completion within 180 days of legislation being signed into law, and that departments and agencies notify the respective committees when this has been accomplished.Appropriators should set aside $20 million in DoD defense-wide O&M funding and add $20 million to DHS’s topline to establish and execute AI mandatory annual training.Mandatory AI Training -- DoD and DHS should require mandatory training designed to improve baseline AI literacy, either online or in person. The training should focus on end users and their ability to collect and manage data, and include a short introduction to AI with an emphasis on machine learning, data management, the capabilities and limitations of AI, software decision-making, probabilistic reasoning, and an introduction to the responsible and ethical development and fielding of AI.59Ideally, these requirements should be integrated into existing user training programs rather creating new, separate training requirements. Annual AI training should be mandatory for five years, followed by an assessment of the need to continue the training.Self-DevelopmentCertified Self-Development_86a1d50c-86a6-11ea-9377-66641c83ea003.9Executive BranchProposed Executive Branch Action -- The DoD, DHS, Department of State (DoS), Department of Commerce (DoC) Bureau of Industry and Security, Federal Bureau of Investigation, and IC should develop a list of approved online courses related to AI. Lists should include at least one course addressing the ethical and responsible use of AI. Funding should come through the Department or Agency’s centralized training or personnel office to allow individuals to directly access funded courses to maximize access. The organizations should authorize a 72-hour pass for uniformed service members who complete online courses, and authorize their use stand-alone or in conjunction with leave, holiday,or weekend periods. They should also authorize eight hours of leave for civilian employees who complete online courses. Both passes and leave should be authorized once per fiscal year quarter.DoDUndersecretary of Defense for Personnel ReadinessWithin the DoD, the Undersecretary of Defense for Personnel Readiness should be responsible for providing guidance on the above recommendations, including the development of an approved list of courses and funding sources.DHSDepartment of State (DoS)Bureau of Industry and SecurityDepartment of Commerce (DoC)Federal Bureau of InvestigationICOnline Course ProvidersA broad array of online courses in data science, data engineering, and AI are readily available to meet individual interests, abilities, and goals (for example, Coursera and Udemy both offer free courses that come from leading experts and universities). Completion of these courses builds familiarity and skills directly related to adopting, integrating, and using AI capabilities inside the government (see Tab 6, Recommendation 1).CourseraUdemyFederal EmployeesRewards can take several forms. Employees can receive reimbursement for course tuition and fees, akin to existing education benefit programs within the Armed Services.Armed ServicesBonuses can be paid for course completion, as is done with special military skill designations or proven proficiency in a foreign language.Enlisted Service MembersFor enlisted service members, points towards promotion can be added, similar to how other resident or non-resident course completions are factored into promotion competitions within some services. Likewise, additional time-off awards can be granted to employees, beyond their normal time accrual, for course completions. Course completions should be documented and taken into account when making assignments.AI CompaniesThere is a legitimate concern that the government might approve online courses of limited technical or educational quality. To avoid this issue, organizations should seek out educational programs from successful AI companies that already offer public or internal education and training and R1 universities (doctoral universities with very high research activity).R1 UniversitiesCongressProposed Legislative Branch Action -- For certified self-development online courses, authorizing language should mandate that the Secretaries of Defense, Homeland Security, State, and Commerce, and the FBI Director, and the Director of National Intelligence each publish a list of approved courses that shall be assessed and revised annually, funding mechanisms, and an incentive system within 120 days of legislation passing, and notification of authorizing committees when the first uniformed service members and civilian full-time employees have begun and completed online courses.Appropriators should set aside $20 million in DoD Personnel funding to establish and execute certified self-development AI online courses and to compensate employees for successful completion. DHS and ODNI should also receive $20 million in total funding each for AI course development and employee compensation. Appropriators should set aside $5 million each to the DoS, DoC, and FBI to establish and execute certified self-development AI online courses and to compensate employees for successful completion.The government should reward and track employees that successfully complete certified AI training and courses. While mandatory training can require government employees to complete courses aimed at providing a minimum baseline knowledge of AI, it will also benefit the government to encourage employees to voluntarily extend their AI literacy beyond those minimum requirements.ProgrammingMeasure and Incentivize Programming Proficiency_86a1db24-86a6-11ea-9377-66641c83ea003.10Executive BranchProposed Executive Branch Action -- The DoD, DHS, IC, and FBI should create tests comparable to the Defense Language Proficiency Test (DLPT) that include coding languages and a second test for AI competency, and expand their use to non-DoD organizations. The DLPT evaluates and scores proficiency on select languages. Service members are given financial rewards for scoring sufficiently high ratings in valued languages. Proficiency is also tracked in service members’ personnel files.62 Departments and agencies should reassess the test annually to ensure that it remains up to date with commercial practices.DoDUnder Secretary of Defense for Personnel and ReadinessWithin the DoD, the Under Secretary of Defense for Personnel and Readiness should be responsible for a coding language testing and incentive pay program for service members.Defense Civilian Personnel Advisory ServiceThe Defense Civilian Personnel Advisory Service should be responsible for extending the program to civilian personnel, including tracking proficiency with the Defense Civilian Personnel Data System.DHSICFBIAI WorkforceNotably, programming ability is only necessary for part of the AI workforce.AI PractitionersMany AI practitioners who focus on algorithm development and the use of existing AI frameworks have minimal programming ability, but are still experts in their field. We recognize that measuring and incentivizing programming ability will improve the government’s software development and some components of the AI lifecycle, but will not be a complete solution on its own.CongressProposed Legislative Branch Action -- Authorizing committees should mandate the development and implementation of an annually reassessed coding language proficiency test and an AI competency test within the DoD, DHS, IC, and FBI. The Armed Services committees should use the FY2021 NDAA to direct that testing and incentive pay be available within the DoD, and for DoD to coordinate with the IC, and DHS within a year of each bill’s passage into law.Appropriate $3 million in DoD O&M funding for the development and implementation of a coding proficiency test. Appropriators should set aside $1.25 million per service and $500,000 for the Office of the Secretary of Defense for coding language incentive pay.The government needs a systematic way to identify employees with existing AI-related skills so that it can put people in the best position to make significant contributions. To do so, the government should measure and incentivize its employees’ computer programming proficiency. Programming proficiency can be incentivized through various means. As with foreign language proficiency, bonuses can be paid for different, quantified levels of comprehension and skill.Programming proficiency should also be documented and taken into account when making assignments.ASVABAdjust the ASVAB to Identify Computational Thinking_86a1e646-86a6-11ea-9377-66641c83ea003.11Under Secretary of Defense for Personnel and ReadinessProposed Executive Branch Action -- The Under Secretary of Defense for Personnel and Readiness should oversee adjustments to the ASVAB to measure computational thinking. Results should be used to place personnel in career fields and to identify personnel for further training, similar to the manner in which the Defense Language Aptitude Battery is used.CongressProposed Legislative Branch Action -- The House and Senate Armed Services committees should use the FY 2021 NDAA to mandate that the Under Secretary of Defense for Personnel and Readiness research and draft a new portion within the ASVAB that identifies and measures computational thinking based on existing academic and private sector testing, with language directing that the addition be incorporated into ASVAB testing a year after the bill is signed into law.The government needs a systematic way to identify potential talent so that it can avoid wasting training resources or putting personnel in positions where they are unlikely to succeed. DoD should adjust the Armed Services Vocational Aptitude Battery Test (ASVAB) to measure computational thinking. We consider computational thinking to include “the thought processes involved in formulating problems and their potential solutions in ways that the solution can beeffectively carried out by an information-processing agent (human or machine, or more generally, by a combination of both). More colloquially, [computational thinking] encompasses a set of processes that defines a problem, breaks it down into components, and develops models to solve the problem, then evaluates the result, iterates changes, and does it again.”63 Once adjusted, thetest should be made available to civilian employees who are interested in voluntarily demonstrating aptitude for additional training.The last major revision of the standardized test was in 2002 (the year most people entering the military in 2020 were born) and mostly involved administrative changes. The test can be updated again to identify computational thinking skills relevant to the military application and operation of information systems, computer language, data science, data engineering, and AI. This includes problem decomposition, abstraction, pattern recognition, analytical ability, the identification of variables involved in data representation, as well as the ability to create algorithms and solution expressions.Student ExposureCreate Opportunities for Students to be Exposed to Government Work by Hiring University Professors as Part-Time Government Researchers_86a1f28a-86a6-11ea-9377-66641c83ea003.12StudentsUniversity ProfessorsGovernment ResearchersSTEM Professor Recruitment ProgramsProposed Executive Branch Action -- The DoE, DoD, IC, DHS, and DoC should establish STEM professor recruitment programs that include:* Outreach to and recruitment at universities, conferences, and professional organizations.* The creation of at least ten part-time researcher billets, five of which are specifically for AI researchers, at laboratories and research facilities including but not limited to the Los Alamos National Laboratory, Oak Ridge National Laboratory, Lawrence Livermore National Laboratory, Army Research Laboratory, Air Force Research Laboratory, Naval Research Laboratory, and the National Security Agency Laboratory for Advanced Cybersecurity Research. Each lab should have the authorization to recruit, screen, select,and compensate part-time professors at the laboratory’s local leadership’s discretion, not through a centralized hiring mechanism at the agency level.* It should be noted that 10 HQEs is a minimum, and laboratories should not interpret the above guidance as a constraint on any other outreach programs or methods for working with faculty.* Specified authorization for research facilities in the program to use direct hiringauthorities and pay in 10 U.S.C. § 1599h. Under section 1599h, the laboratories of the military departments, DARPA, and Director of Operational Test and Evaluation have the authority to hire experts in science and engineering under their own program of personnel management authority. This authority exempts authorized programs from some sections of title 5 that address hiring processes and terms of employment, including allowing for 150 percent of the vice-president’s salary in financial compensation in a single fiscal year. This would allow organizations that fall under section 1599h to hire professors in part time capacities at a competitive salary.* Specific authorization for research facilities to hire experts, consultants, and highly qualified experts outside of the competitive service system. Section 3109 of Title 5 authorizes agencies to hire experts whose education and experience allows them to perform tasks beyond the range of a competent person in that field and consultants that can provide valuable and pertinent advice. Both types of employee can serve up to one year or on an intermittent basis, and their base pay rate cannot exceed that of a GS-15 step 10’s hourly rate.67 Another section of Title 5 authorizes the DoD to hire up to 2,500 highly qualified experts for up to five years, with an additional year at the Secretary of Defense’s discretion.68* A referral bonus for other faculty hired under the part-time faculty researcher program and interns who participate in laboratory internship programs and the Pathways Internship Program.* Expand the use of Cooperative R&D Agreements (CRADAs) between government organizations, universities, and the private sector to enable joint research and expertise sharing.69DoEDoDICDHSDoCLos Alamos National LaboratoryOak Ridge National LaboratoryLawrence Livermore National LaboratoryArmy Research LaboratoryAir Force Research LaboratoryNaval Research LaboratoryNational Security Agency Laboratory for Advanced Cybersecurity ResearchCongressProposed Legislative Branch Action -- Authorizing committees should mandate that the DoE, DoD, DoC, DHS, and IC begin programs to hire university faculty with relevant STEM expertise as part time researchers at Los Alamos National Laboratory, Oak Ridge National Laboratory, Lawrence Livermore National Laboratory, Army Research Laboratory, Air Force Research Laboratory, Naval Research Laboratory, National Security Agency Laboratory for Advanced Cybersecurity Research, and other laboratories at their discretion within 180 days of the passage of legislation. Performance measures include:* the number of faculty hired within 365 days of the passage of legislation;* after two years, the number of college interns and recently graduates interviewed and/or hired after a reference from faculty part-time researchers each year; and* the results of faculty part-time researchers work after two years. ^Appropriators should set aside such sums as necessary for the DoE to hire 10 university professors as part-time researchers at each DoE laboratory, up to a total of 170 professors, and to award recruitment bonuses. Appropriators should set aside such sums as necessary for the DoD to hire 10 university professors as part-time researchers at each DoD laboratory, up to 100 professors, and to award recruitment bonuses. Appropriators should set aside such sums asnecessary for the IC to hire 10 university professors as part-time researchers at each IC laboratory, up to a total of 30 professors, and to award recruitment bonuses. Appropriators should appropriate such sums as necessary for the DoC to hire 10 university professors as part-time researchers at each DoC laboratory, up to a total of 30 professors, and to award recruitment bonuses.To create opportunities to interact with students, some private sector companies hire universityfaculty as summer or part-time researchers. The companies benefit from access to a diverse groupof experts that understands and often creates the world’s most cutting-edge AI. In turn the companies provide resources, exposure to new techniques, and financial compensation to the professors, sometimes including funding for their university-based lab. When the professors return to teaching, they also expose promising students to the companies’ work, creating student awareness and excitement about the available opportunities, a positive perception of the companies, and relationships that encourage student employment upon graduation.To replicate this proven technique, the Department of Energy (DoE), DoD, DoC, DHS, and IC should hire university faculty with relevant STEM expertise to serve as part-time researchers in government laboratories. Faculty members could work during sabbaticals, summer breaks, or limited hours throughout the year.Some government laboratories already hire university professors as part-time researchers. It is unclear how often this happens, the return on investment when it takes place, or how often faculty expose their students to government work, internships, or employment. These metrics should be tracked and, if helpful, the programs should expand and specifically include faculty with expertise in AI.InternshipsIncrease the Use and Utility of Pathways Internships_86a1f97e-86a6-11ea-9377-66641c83ea003.13Pathways Internship ProgramProposed Executive Branch Action -- The Office of Personnel Management (OPM) should expand the Pathways Internship Program.OPMOPM should change the requirements for computer science, mathematics, electrical engineering, and computer engineering majors to allow conversion within 365 days of completing a program, and after completing 220 hours of work.STEM StudentsThis will allow students who complete an internship the summer between their junior and senior years to initiate the security clearance process, completetheir senior year, and convert to a full-time position after graduation.National Security Departments & AgenciesNational security departments and agencies should increase their student recruitment efforts, with a focus on recruiting undergraduate and graduate students with computer science, mathematics, electrical engineering, and computer engineering majors.Departments and agencies participating in the Pathways Internship Program should eliminate the application cap and be required to develop their own position-specific qualification standards for Pathways Internship positions.National Security Departments & AgenciesThree major challenges currently reduce the Pathways Internship Program’s effectiveness as a hiring mechanism. The first is a cap on the number of qualified applicants. Today, some government internship program application windows are only open for minutes before they reach the limit on the number of applications and close. This prevents well-qualified applicants from participating in the program. National security departments and agencies should eliminate such caps on the number of applicants.Government AgenciesThe second challenge is the OPM Qualification Standards. Federal regulations allow individuals to “be evaluated against either agency-developed standards or the OPM Qualification Standard for the position being filled.”71 Many agencies default to the generic OPM standards72 rather than creating their own, position-specific standards. Government agencies should create their own, position-specific qualification standards for internship positions.Pathways InternsThe third challenge is the difficulty many Pathways Interns face when trying to convert to a full-time, competitive service position upon completion of their educational requirements. Currently, internships may be converted to a permanent position within 120 days of completing the program if they have completed 640 hours, completed their degree, meet the position’s qualifications and requirements, and perform their job successfully.73 This structure makes it difficult for students to convert a summer internship into a full-time position the following year.CongressProposed Legislative Branch Action -- Authorizing committees should require the DoD, IC, DoS, DoC, DHS, and Department of Justice to increase the number of Pathways Internships and the number of conversions to full-time positions, including for mathematics and computer science majors, within 365 days of thepassage of legislation.Authorizing committees should direct the Office of Personnel Management to change 5 CFR § 362.203(c) to read: “Qualifications. Individuals will be evaluated against agency-developed standards.”Authorizing committees should direct the Office of Personnel Management to change 5 CFR § 362.204(b)(1) to read: “Completed 220 hours of work experience acquired through the internship program, except as provided in paragraphs (c) and (d) of this section, while enrolled as a full-time or part-time, degree- or certificate-seeking student.”Authorizing committees should direct the Office of Personnel Management to change 5 CFR § 362.204(b)(2) to read: “Completed a course of academic study, within the 365-day period preceding the appointment, at a qualifying educational institution conferring a diploma, certificate, or degree.”Performance metrics Congress should monitor include:* The change in the number of Pathways Internship Program participants from the three previous fiscal years to the current fiscal year,* The change in the number of mathematics and computer science majors participating in the Pathways Internship Program from the three previous fiscal years to the current fiscal year,* The change in and overall number of conversions from the Pathways Internship Program to full-time positions from the three previous fiscal years to the current fiscal year, and* The change in the number of mathematics and computer science majors who convert from the Pathways Internship Program to full-time positions from the three previous fiscal years to the current fiscal year.The government should increase the use of the Pathways Internship Program as a mechanism to screen and hire new employees. Many government organizations already use internships as a hiring pipeline. Effective internship programs allow agencies to vet candidates, begin the security clearance process early enough for students to receive a clearance prior to graduation, establish relationships with students before they commit to another employer, and to expose students to the benefits of government work.70CyberCorpsExpand the CyberCorps: Scholarship for Service_86a20478-86a6-11ea-9377-66641c83ea003.14CyberCorpsCyberCorpsProposed Executive Branch Action -- The National Science Foundation and Office of Personnel Management should broaden the CyberCorps: SFS to pay for up to four years and to include fields falling under digital engineering, as those fields are defined by the National Defense Authorization Act for Fiscal Year 2020 (Pub. L. 116-92, section 230): “the discipline and set of skills involved in the creation, processing, transmission, integration, and storage of digital data, including data science, machine learning, software engineering, software product management, and artificial intelligence product management.”National Science FoundationOffice of Personnel ManagementCongressProposed Legislative Branch Action -- The Armed Services committees should broaden the CyberCorps: SFS to include digital engineers, as defined by section 230 of the National Defense Authorization Act for Fiscal Year 2020, to pay for up to four years of scholarships, and to include the opportunity to begin the security clearance process.The Armed Services committees should amend 15 U.S.C. § 7442 subsection (a) to read: “...recruit and train the next generation of information technology professionals, digital engineers, artificial intelligence practitioners, data engineers, data analysts, data scientists, industrial control system security professionals, security managers, and cybersecurity course instructors to meet the needs of the cybersecurity mission for Federal, State, local, tribal, and territorial governments.”75The Armed Services committees should amend 15 U.S.C. § 7442 subsection (b) to provide an opportunity for scholarship recipients to initiate their security clearance process at least one year before their planned graduation date.The Armed Services committees should amend 15 U.S.C. § 7442 subsection (c) to allow the scholarship to last for 4 years.The Office of Personnel Management and National Science Foundation should expand the scopeof CyberCorps: Scholarship for Service (SFS). The CyberCorps SFS is a recruiting programdesigned to attract students studying IT, cybersecurity, and related fields into the USG.Expanding it could bring in more people with AI-related skills. It is managed by the NationalScience Foundation in partnership with the Office of Personnel Management and the Department of Homeland Security. Students enrolled in the program receive a scholarship in exchange for anobligation to work in an approved government agency for a period of time equal to the timecovered by the scholarship. Seventy undergraduate and graduate institutions participate in SFSby selecting students for the program, and since 2001, 3,600 students have received scholarships,94 percent of whom went on to serve in government. Hiring typically takes place during annualonline and in-person career fairs.74It should be noted that cyber and AI are different fields. Expanding CyberCorps: SFS toCyberCorps and AI: SFS would avoid increasing administrative burdens. This should not be taken as an indication that AI and cyber are synonymous, as the education and skills for each field differ.Also, the National Security Commission on AI believes the United States Government and Department of Defense may benefit from developing a digital corps, and is considering the steps needed to create one. The above recommendation is consistent with that end, and can be easily incorporated into a digital corps if the decision to pursue one is approved.Fellowships & PartnershipsIncrease the Number of Fellowships and Partnerships with Industry, and Increase the Number Focused on Artificial Intelligence and Software Development_86a210ee-86a6-11ea-9377-66641c83ea003.15IndustryOffice of the Secretary of DefenseSecretary of Defense Executive Fellow Program: The Office of the Secretary ofDefense should expand the Secretary of Defense Executive Fellow Program. In addition to previous nomination requirements, each service and the Office of the Secretary of Defense should be required to nominate an additional eight uniformed service members or government civilians for sponsorship by major corporations that are leading in AI applications.Undersecretary of Defense for Acquisition and SustainmentPublic-Private Talent Exchange Program: The Undersecretary of Defense forAcquisition and Sustainment’s Public-Private Talent Exchange Program should both expand and increase its partnership with AI focused companies. Without changing the private sector partners of the current five employee cohort, the Undersecretary of Defense for Acquisition and Sustainment should grow the program by ten government employees who work with AI companies and ten private sector participants from AI focused companies.AI Focused CompaniesUndersecretary of Defense for Research and EngineeringThe Undersecretary of Defense for Research and Engineering, Department of Defense Chief Information Officer, military service secretariats, andOffice of the Director of National Intelligence should establish similar AI-centric, 20-person talent exchange programs.Department of Defense Chief Information OfficerMilitary Service SecretariatsOffice of the Director of National IntelligenceUSAFUSAF Education with Industry Program: The Air Force’s Education with IndustryProgram should be replicated by the other military services and the IC, but with a focus on private sector companies working on AI applications. Likewise, participation in these programs should be treated, for the purposes of promotion boards and subsequent assignments, as equivalent to attending resident professional military education.DoDPublic-Private Exchange Program Billet Office: DoD should establish a PublicPrivate Exchange Program Billet Office whose primary purpose is to temporarily hold billets for civilian DoD participants in exchange programs, roughly comparable to student detachments for active-duty personnel attending joint professional military education programs. Currently, DoD employing offices are disincentivized from encouraging employees to participate in exchange programs, as they lose a member of their team for an extended period, often without a replacement due to the strictness of the billeting process.79 A Public Private Exchange Program Billet Office can reduce this disincentive by holding temporary billets for employees that participate in exchange programs, allowing employing offices to temporarily backfill their loss. To avoid the creation of yet another disincentive, employees who participate in talent exchange programs should be guaranteed that their previous position will remain available, even ifit is temporarily filled while they participate in the exchange program.CongressProposed Legislative Branch Action -- Authorizing committees should amend 10 U.S.C. § 1599g to direct the DoD and Office of the Director of National Intelligence to exchange talent and promote training with the commercial tech sector working on AI applications. Section 1599g should also reflect minimum participation requirements for talent exchange programs, as outlined above. The Office of the Secretary of Defense, military service secretaries, and Office of the Director of National Intelligence should provide an annual report to the authorizing committees about the status of their talent exchange programs.National security departments and agencies should increase the number of fellowships andpartnerships with industry, particularly fellowships with AI and software companies. The numberof fellows, particularly from the DoD, needs to increase significantly if service members andcivilian employees are to garner and leverage the AI-related experience and expertise resident inthe private sector. Moreover, individual participation in these fellowships should be treated, forthe purposes of promotion boards and subsequent assignments, as having an experienceequivalent to graduating from a resident professional military education program.Notably, this is not a recommendation to shift the focus of fellowships and partnerships withindustry away from other fields, many of which add significant value to the government. Instead,the number of government and private sector employees participating in fellowships andpartnerships with industry should significantly increase, with many of the increased number goingto or coming from AI and software companies.Hardware & 5GPromote U.S. Leadership in AI Hardware & 5G_86a21698-86a6-11ea-9377-66641c83ea004HuaweiQ1 recommendations lay the groundwork for long-term access to resilient, trusted, and assured microelectronics for AI advantage, and takes a portfolio-based approach to ensure that the United States continues running faster than potential adversaries in the field of cutting-edge microelectronics. The Commission recommends expanding USG AI-enabling microelectronics programs to develop novel and resilient sources for producing, integrating, assembling, and testing AI-enabling microelectronics; stating research priorities, increasing USG R&D funding, and articulating a national strategy for microelectronics and associated infrastructure in order to maintain global leadership in AI-enabling hardware. The Q1 memo also provides limited, near-term recommendations to bolster U.S. fifth-generation cellular wireless (5G) capabilities, including by recommending policies and funding opportunities that would advance spectrum-sharing and 5G commercial licensing, support R&D in critical technical areas, and further the development of open-access radio networks, in order to accelerate 5G adoption in the United States and foster global alternatives to Huawei.Microelectronics ProgramsExpand USG AI-Enabling Microelectronics Programs_86a21cce-86a6-11ea-9377-66641c83ea004.1Expand USG access to high-performance, secure microelectronic components by developing novel and resilient sources for producing, integrating, assembling, and testing AI-enabling microelectronics.Packaging, Assembly & TestingExpand the advanced packaging, assembly and testing technical execution area under the Trusted and Assured Microelectronics Program_86a22958-86a6-11ea-9377-66641c83ea004.1-1NavyExpand the Navy led, DoD-wide existing advanced packaging, assembly and testing technical execution area under the Trusted and AssuredMicroelectronics Program that is led by the Office of the Under Secretary of Defense for Research and Engineering. Create a trusted state-of-the-art AI hardware demonstration prototype for multi-chip packages (AI-MCP) by reprogramming $50 million in FY 2020 or appropriating $50 million in FY 2021.Office of the Under Secretary of Defense for Research and EngineeringTrusted and Assured Microelectronics ProgramIn FY 2020, DoD is pursuing public-private partnerships under its Trusted and Assured Microelectronics Program to develop heterogeneous integration and advanced packaging of microelectronics in a secure U.S. manufacturing facility. The existing program does not include an AI specific application prototype demonstrator. This recommendation would expand the program to incorporate AI-specific capabilities for reduced size, weight, and power (SWaP), with increased performance and IP protection necessary for forward deployed DoD weapon system applications.Naval Surface Warfare Center CraneAs part of the Trusted and Assured Microelectronics program, Naval Surface Warfare Center Crane (NSWC Crane) initiated a SOTA Heterogeneous Integrated Packaging (SHIP) prototype program utilizing an Other Transaction Authority (OTA) agreement with a U.S.-based global leader in microelectronics in FY 2020.84 With an additional $50 million, SHIP could expand the existing pilot prototype program to include heterogeneous integration of multi-chip packages incorporating AI specific chips and configurations. SOTA digital processing chips could include Field Programmable Gate Arrays (FPGAs), Application Specific Integrated Circuits (ASICs), or Graphics Processing Units (GPUs) tightly coupled with memory and leading edge, high-speed communications chips to reduce processing latency and also include “root-of-trust” security features to protect IP.Defense Industrial Base PartnersSHIP would partner with one or more defense industrial base partners in conjunction with DoD program offices that would be responsible for transitioning and fielding the AI-MCP capability once successfully demonstrated.DoD Program OfficesDoD ComponentsA successful pilot program would also allow any DoD component—and potentially any USG organization—to purchase the AI-enabled microelectronics MCP using the same contract vehicle.This funding could be reprogrammed in FY 2020 or included in DoD’s FY 2021 appropriation.Without the additional $50 million in FY 2021 funding, AI-centric capabilities will not be included in the pilot prototype program, delaying the development of heterogeneously integrated AI-specific microelectronics capabilities indefinitely.Semiconductor ManufacturingFund an accelerated site survey for a U.S.-based commercial front-end-of-line semiconductor manufacturing facility._86a22f2a-86a6-11ea-9377-66641c83ea004.1-2Office of the Director of National IntelligenceFund an accelerated site survey by the Office of the Director of National Intelligence for a U.S.-based commercial front-end-of-line semiconductor manufacturing facility by reprogramming funding in FY 2020 funding or by appropriating funding in FY 2021.Institute for Defense AnalysisAt present, the USG does not have trusted access to SOTA microelectronics manufacturing. Numerous studies conducted by or on behalf of the U.S. government have highlighted this issue and offered recommendations, including work by the Institute for Defense Analysis, the National Academy of Sciences, and DoD’s Office of Military Industrial Base Policy, among others.85 However, many well-founded recommendations have not been adopted and the challenge has only compounded with time.National Academy of SciencesSemiconductor CompaniesThe Office of the Director of National Intelligence has assessed opportunities to partner with U.S.-owned private-sector semiconductor companies to establish a SOTA semiconductor design, fabrication, and packaging facility in the United States that would meet USG technical and security needs for high-end semiconductors and help achieve cutting-edge performance levels.SHIP ProgramThis capability would complement DoD’s SHIP program.Private Sector FirmsWhile building a cutting-edge, high-capacity semiconductor fabrication plant for dedicated-government use would likely cost approximately $20 billion, the ODNI approach calls for a security-based split-manufacturing facility and partnering with a private sector firm to build a facility, which would produce both commercial-use and government-use chips. The Front-End-of-Line (FEOL) portion—also known as circuit building blocks—could be produced within the standard commercial areas of the facility, along with standard commercial chips. For chips that will be used for controlled but unclassified, export-controlled, or classified uses, the Back-End-of-Line (interconnects and circuit function) layers could be produced on a separate government-certified secure line either in the same facility or at a separate facility, to which access could be controlled. The final assembly, packaging, and testing would also occur in a secure facility. This arrangement provides additional supply chain integrity and protects sensitive USG IP.86The ability to reliably source trusted state-of-the-art microelectronic components is critical for United States commercial companies and for current and future USG national security capabilities...For economic competitiveness and national security, the USG must act swiftly to meet the challenge of assured access to SOTA microelectronics manufacturing while simultaneously protecting sensitive IP and ensuring that manufactured parts operate only as intended...The site survey is an essential first step toward moving forward with this project. This step can be funded and started immediately, would provide a specific cost of the overall project within 12-18 months, and could have the effect of lowering the overall estimated cost of the facility. The FY 2021 President’s Budget request for microelectronics does not include funding for this site survey. Adding funding for the survey in FY 2021 would ensure that facility construction couldproceed as quickly and efficiently as possible, should the USG decide to move forward with this recommendation.FundingContinue fully funding DoD’s Trusted and Assured Microelectronics Program in the FY 2021 Budget._86a23b46-86a6-11ea-9377-66641c83ea004.1-3CongressIn the FY 2021 President’s Budget, DoD’s Trusted and Assured Microelectronics research, development, test, and evaluation (RDT&E) request includes $489 million for advanced component development and prototypes (ACD&P) and $108 million for system development and demonstration (SDD).87 Combined, these programs will improve access to advanced packaging and testing; support the development of quantifiable assurance and secure design; develop foundry access standards; expand access to non-complementary metal oxide semiconductor (CMOS) SOTA microelectronics; support disruptive research and development; and promote education and workforce development. These are foundational microelectronics capabilities that will also enable the development and application of AI/ML capabilities across national security mission areas. Congress has fully funded the Administration’s request for the DoD Trusted and Assured Microelectronics Program in each of the past three years. This recommendation endorses the Administration’s request for Trusted and Assured Microelectronics in FY 2021.TransitioningDevelop clear metrics for transitioning AI-enabling microelectronics from research programs to operating forces and the commercial sector._86a2456e-86a6-11ea-9377-66641c83ea004.1-4OUSD/R&EOUSD/R&E should develop clear metrics for transitioning AI-enabling microelectronics from research programs to operating forces and the commercial sector.DoDAI-specific microelectronics could accelerate the performance of machine learning capabilities across DoD mission areas, but only if they can be manufactured at scale. At the same time, more intellectual property generated through DoD research programs could be transitioned to the commercialsector to further spur innovation.As a first step toward these goals, DoD should clarify its approach and plans for transitioning research efforts on advanced microelectronics to battlefield advantage through existing programs of record. It should also identify metrics for transitioning intellectual property it does not need to exclusively retain to enable further research and development outside the government, especially in the commercial sector.USD R&EThis recommendation urges USD R&E to develop specific metrics for transitioning custom and semi-custom advanced microelectronics (e.g., GPUs, FPGA, and ASICs) to programs of record.DoD R&E EnterpriseThese metrics should capture the transition plans and milestones for existing microelectronics research programs across the DoD R&E enterprise, including the Defense Advanced Research Projects Agency's (DARPA) Electronics Resurgence Initiative and the Trusted & Assured Microelectronics program. It should also feature narrative descriptions of the impact to mission of successful transitions. Transition metrics should also identify milestones for scaling themanufacturing of SOTA, AI-enabling electronic components developed through research efforts to meet the needs of DoD programs.DARPACommercial SectorFor commercialization, metrics should track the success of intellectual property sharing with the commercial sector for microelectronics programs.Despite successful research efforts to advance the fields of microelectronics and AI, pathways for transitioning advanced capabilities to the battlefield remain challenging.LeadershipMaintain global leadership in microelectronics by clearly stating research priorities, increasing USG R&D funding, and articulating a national strategy for microelectronics and associated infrastructure (e.g., a national microelectronics laboratory and incubator)._86a24c3a-86a6-11ea-9377-66641c83ea004.2Beyond-CMOS HardwarePrioritize research into beyond-CMOS AI hardware capabilities for national security applications._86a25392-86a6-11ea-9377-66641c83ea004.2-1Executive Branch AgenciesTo complement ongoing efforts, the Commission wishes to highlight for Executive Branch agencies and Congress specific, beyond-CMOS hardware capabilities that the Commission believes will be essential to ensure the United States maintains its current edge in advanced AI-related hardware.DARPAIn particular, the Commission believes that DARPA and the NSF should prioritize programs which focus on research involving 3D chip stacking, photonics, carbon nanotubes, Gallium Nitride (GaN) transistors, domain specific hardware architecture, and electronic design automation, and cryogenic computing. These represent technologies which the Commission assesses are key to ensuring the continued growth of compute and mitigate the impacts of the slowdown of Moore’s Law. While there has been limited USG-funded research on these topics, USG funding has been heavily weighted towards applied research, rather than the basic research that is necessary to achieve breakthroughs.NSFThe USG must prioritize R&D into technologies and techniques that can extend the life of classical computing, ensuring continued growth in processing power until potential alternative computing architectures such as quantum or neuromorphic computing are readily available.Doing so will be key to ensuring that the United States continues to out-innovate its competitors, particularly as China continues to invest heavily in its domestic semiconductor industry (see Appendix for background).Due to the high costs associated with scaling power and frequency in microelectronics, industry is pursuing performance gains—particularly for Ai—through chip specialization. The result has been specialized GPUs, TPUs, ASICs, or FPGAs which are more efficient at processing AI algorithms than traditional CPUs. While these chips have achieved notable performance gains, they will likely not be a primary driver of long-term performance growth. Specialization faces diminishing marginal returns over time; if advancement of the underlying microprocessors slows, future gains from increased specialization alone will be insufficient to spur significant performance increases. This may lead to market failures, as firms may not see it in their short-term interest to invest in the R&D necessary to lay the foundation for long-term advancementsin microelectronics.89 As a result, government investment is required to spur long-term research and development.Electronics ResurgenceIncrease the budget of DARPA’s Electronics Resurgence Initiative to $500 million._86a25e5a-86a6-11ea-9377-66641c83ea004.2-2Electronics Resurgence Initiative (ERI)The Electronics Resurgence Initiative (ERI) is DARPA’s primary mechanism for funding advanced microelectronics research, investing approximately $250-300 million annually. Much of this program specifically targets beyond-CMOS architectures, funding both applied sciences and basic research at approximately a 5:1 ratio.DARPASince FY 2019, DARPA’s funding request for ERI-related programs has declined. Meanwhile, the Administration’s FY 2021 overall budget request for DARPA included a 13 percent increase over FY 2019 levels.90 Given the need to achieve microelectronics breakthroughs to ensure continued and long-term leadership in AI, particularly in the area of advanced fabrics and post-CMOS architectures, these two funding streams must move in tandem with one another, not in opposite directions.Defense Science Board (DSB)DARPA’s work on advanced microelectronics is both underfunded relative to the magnitude of the problem set, and currently trending in the wrong direction. In 2018, the Defense Science Board (DSB) recommended that DARPA increase funding for ERI to $450 million to accelerate the discovery of advanced microelectronics fabrics.91 This option would endorse and build on the DSB recommendation, recommending that the budget for ERI increase to $500 million in FY 2021.Microelectronics & SemiconductorsIncrease NSF’s topline budget by $50 million directed toward microelectronics and semiconductors research for enhancing AI capabilities._86a2671a-86a6-11ea-9377-66641c83ea004.2-3National Science FoundationThe National Science Foundation’s budget for semiconductors and microelectronics research has declined approximately 10 percent since FY 2019, from $92.59 million to $84.16 million in the FY 2021 budget request.93 Meanwhile, NSF’s FY 2021 request for AI represents an 81 percent increase over the FY 2019 allocated funding. While the NSF’s AI-related funding increase has been highlighted, the decrease in semiconductors and microelectronics R&D is a misstep given the critical relationship between leadership in AI-related hardware R&D and leadership in AI writ large. AI applications are inherently reliant on semiconductors and microelectronics for their functionality, and research into both advanced microelectronics fabrics and technologies relevant to specialized AI-related chipsets will enable significant future performance gains and AI capabilities.CongressCongress should appropriate an additional $50 million to the NSF FY 2021 budget and highlight the importance of research into semiconductors and microelectronics targeted to enhance AI capabilities.94 This increased funding level would reverse the decline in funding for basic research in advanced microelectronics design techniques.AI-Enabled HardwareFund a pilot $20 million prize challenge for AI-enabled hardware through IARPA._86a26ddc-86a6-11ea-9377-66641c83ea004.2-4IARPAIntelligence Advanced Research Projects Activity (IARPA) currently has no programs focused on advanced chip design after concluding the Trusted Integrated Circuits (TIC)95 and Rapid Analysis of Various Emerging Nanoelectronics (RAVEN)96 programs, both of which focused on microelectronics. This is a notable gap, given IARPA’s history of successful programs in this area. IARPA could leverage its successful history of running “prize challenges” to invite industry and academia to attempt breakthrough high-risk, high-payoff research. Drawing on this track record, IARPA could establish a new prize challenge focused on advanced chip design and AI-enabling microelectronics. This challenge, or series of challenges, could attempt to spur innovation towards specific goals, such as developing secure electronic design automation (EDA) libraries, or shrinking the time to transition from an algorithm to an ASIC in one year.CongressCongress should appropriate $20 million to IARPA in FY 2021 to run a pilot program creating prize challenges specifically associated with advancing AI-enabled hardware, with the goal of spurring advances in scaling post-Moore’s law. Should this prove successful and scalable, IARPA can evaluate whether expanding the program by adding additional resources in future years would prove beneficial.Microelectronics StrategyDevelop a national microelectronics strategy._86a279b2-86a6-11ea-9377-66641c83ea004.2-5Department of CommerceGiven the urgent and cross-cutting nature of the problem, it is critical that DoD’s approach be integrated with actions taken by Commerce and State, as well as other relevant agencies, to promote domestic R&D of microelectronics and prevent the illicit transfer of technology to competitors.Department of StatePublic SectorThe United States also faces infrastructure and funding gaps for microelectronics research, development, and commercialization in the public and private sectors, all of which are important for maintaining U.S. leadership in semiconductors.Private SectorThe United States must continue innovating at the cutting edge of semiconductors to strengthen the U.S. economy and mitigate the threat posed by Chinese industrial policy and corresponding investments.98 A U.S. strategy would foster a coordinated approach and help overcome looming challenges to microelectronics innovation, competitiveness, and supply chain integrity.DoDDoD, in coordination with other relevant departments and agencies including the DOE, ODNI, the Department of State, the Department of Commerce, and NSF, could develop the strategy within six months of the passage of the NDAA, drawing on the findings of the 2017 President's Council of Advisors on Science and Technology report on semiconductor leadership,99 the 2019 DSB’s Summer Study, and DoD’s existing strategy for assured access to trusted microelectronics.100 This strategy would build on previous studies and strategies by focusing on AI-enabling performance, research fundingpriorities, interagency and international coordination, and private sector partnerships.DOEODNINSFSenior Microelectronics Leadership Panel The Commission also recommends establishing a senior leadership panel composed of the National Security Council, the National Economic Council, the Office of Science and Technology Policy, and relevant departments and agencies within 30 days to oversee U.S. microelectronics leadership and competitiveness.National Security CouncilNational Economic CouncilOffice of Science and Technology PolicyCongressThe legislative language should also require that the Secretary of Defense and Assistant to the President for National Security Affairs provide an interim briefing to Congress after 90 days regarding the strategy’s status, and report any initial findings and conclusions.Assistant to the President for National Security AffairsSecretary of DefenseAs part of this strategy, the Secretary of Defense—in coordination with the Secretary of Energy and Director of National Intelligence—should consider the viability, efficacy, and cost of developing a national laboratory exclusively focused on microelectronics research and development.Secretary of EnergyDirector of National IntelligenceNational Microelectronics LaboratorySuch a laboratory could serve as a hub for federal research into breakthrough microelectronics-related technologies,101 as well as an incubator for early-stage semiconductor startups which currently face difficulties scaling due to the high costs of microelectronics design and fabrication.102Semiconductor StartupsMicroelectronics StartupsThe incubator could provide resources to promising, early-stage microelectronics startups, while also giving them access to fabrication facilities, design tools, and shared intellectual property to assist with early-stage development costs. This laboratory could also become a center for government expertise in high-performing, trusted microelectronics, ensuring continued U.S. leadership in this field over the ensuing years.DiplomatsThe national strategy should also consider the role of diplomacy and trade in maintaining U.S. microelectronics leadership.U.S. AlliesIn particular, it should assess the potential impact of approaches to multilateral export controls tailored for specific semiconductor manufacturing equipment (e.g., extreme ultraviolet photolithography equipment and argon fluoride immersion photolithography equipment) through direct coordination with key U.S. allies, the Wassenaar Arrangement, and other multilateral fora.103ChinaIt should also analyze if additional U.S. trade enforcement actions are necessary in light of China’s extensive semiconductor subsidy program.U.S. Semiconductor IndustryFinally, it should assess options for increasing the competitiveness of the U.S. semiconductor industry through the elimination of policies that harm U.S. companies without producing a substantial benefit to U.S. national security (e.g., trade barriers and uniliteral export controls).Require the USG to develop a national microelectronics strategy within 180 days and assess the viability of a national microelectronics laboratory and incubator.The United States lacks a national strategy for microelectronics to support interagency coordination within the USG externally with industry and academia. A truly national strategy would build on DoD’s Microelectronics Innovation for National Security and Economic Competitiveness program and previous studies conducted by the U.S. Government or on its behalf. Expanding beyond DoD, it would leverage the expertise and perspectives of the interagency to produce a fully coordinated approach to microelectronics.97 ...The Commission recommends the FY 2021 NDAA include language requiring the creation of a national strategy with components focused on national security, U.S. leadership, and competitiveness, within 180 days. 5GPursue a portfolio approach to accelerating 5G adoption in the United States and fostering global alternatives to Huawei._86a27ffc-86a6-11ea-9377-66641c83ea004.3HuaweiPursue policies and funding opportunities which advance spectrum-sharing and 5G commercial licensing, as well as R&D into 5G software, hardware, microprocessing technology, and open-access radio networks, as part of a portfolio approach to accelerating 5G adoption in the United States and fostering global alternatives to Huawei.ResourcesAppropriately resource R&D efforts to develop and improve 5G spectrum sharing._86a286dc-86a6-11ea-9377-66641c83ea004.3-1AI Systems ProvidersAI systems require high fidelity sensing as well as fast, safe, and secure networks.U.S. MilitaryIt is a national security imperative for the U.S. military to have access to a powerful 5G network to enable future AI capabilities, and ensure the network is trusted to prevent competitors from accessing our AI systems. The United States must preserve this access and trust while building its 5G network.DoDThe sub-6 GHz spectrum, sometimes referred to as mid-band or the goldilocks band of spectrum, is a critical portion of the spectrum for both DoD and commercial 5G operations. It is currently used by many radar and communication systems within the DoD, as radars and sensors operatingin the sub-6 GHz region of spectrum combine high discrimination capability with long range operations.Civilian Communication Service ProvidersSub-6 GHz is also critical for 5G civilian communications since it provides high data rates together with good range and penetration.Given that sub-6 GHz is important for both sensing using radar and communications, spectrum sharing research is particularly important to enable access for both purposes simultaneously.However, current spectrum sharing capabilities are not optimized.104 Additional technical development on spectrum sharing, particularly in utilizing AI to dynamically manage spectrum access, is still necessary to fully utilize the available spectrum for sharing. Implementation of improved spectrum sharing technologies could make more mid-band spectrum available for civilian and commercial use while preserving critical DoD systems for sensing.R&D efforts should be appropriately resourced to develop and improve 5G spectrum sharing, particularly in the sub-6 GHz mid-band spectrum...The Commission supports increased research and development to improve the U.S. government’s ability to share the sub-6 GHz mid-band spectrum with 5G operators for commercial and civilian use.Sub-6 GHz SpectrumExpand sub-6 GHz spectrum-sharing arrangements and licenses for commercial 5G._86a3c5d8-86a6-11ea-9377-66641c83ea004.3-2ChinaThe slow rollout of 5G networks in the United States, which currently has one-fifteenth of the number of deployed 5G-operable base stations compared to China,105 risks slowing U.S advances in AI, both in the government and the private sector.106Commercial Spectrum UsersThe lack of U.S. mid-band spectrum commercial availability, combined with the high capital requirements necessary to ensure coverage utilizing high-band spectrum, have substantially slowed the deployment of 5G networks.Other CountriesFurthermore, the high-band is losing out to mid-band as the global standard, as other countries face less spectrum-crowding in the mid-band and are more willing to adopt it as their standard.HuaweiThis trend is exacerbated by Huawei’s emphasis on the mid-band, and the $75 billion that the Chinese government has given it in subsidies.107Citizens Broadband Radio ServiceThe USG is working to address this problem by developing spectrum sharing capabilities within the 3 to 6-GHz range.FCCIn 2015, the FCC established the Citizens Broadband Radio Service (CBRS), the first U.S. spectrum sharing model.NTIASince that time, the NTIA has studied, and has collaborated with the DoD and FCC, on maximizing spectrum sharing capabilities.108 The CBRS enables shared federal and non-federal use of the band.U.S. NavyThis work will enable the U.S. Navy and non-government providers to share the 3550-3700 MHz band, and will be divided into three, dynamically-managed tiers: the Navy will maintain first priority access, followed by companies and organizations which purchase priority-access licenses, and finally companies and organizations that register at no cost.109 The FCC will hold an auction for priority-access licenses for this band in July 2020.110Non-Government ProvidersDoDExpanding such sharing efforts is critical, as it will ensure that DoD maintains access to spectrum which is essential for operational effectiveness, while also broadening commercial access to spectrum which can be utilized for 5G networks.Defense Innovation BoardIn April 2019, the Defense Innovation Board issued a report which argued that the status quo of spectrum allocation is unsustainable and DoD must expand its sub-6GHz spectrum sharing operations to enable the United States to compete with China in 5G.111Urge the FCC, DoD, and NTIA to work to expand sub-6 GHz spectrum-sharing arrangements and licenses for commercial 5G use, which minimizes risk to DoD operations while expanding the availability of commercial 5G bandwidth...The Commission urges the FCC, DoD, and NTIA to expand spectrum-sharing programs such as the CBRS, and work to license additional sub-6GHz spectrum to wireless carriers and equipment makers for commercial 5G use. Sharing and licensing additional mid-band spectrum will ensure unrestricted DoD access in the event of an emergency. By expanding commercial access to the mid-band, it will also allow the United States to more quickly develop its domestic 5G networks and develop ways and means to compete with Huawei globally.Telecommunications ActPass the Utilizing Strategic Allied (USA) Telecommunications Act._86a3ced4-86a6-11ea-9377-66641c83ea004.3-3CongressHuaweiCombined, these R&D efforts would also directly counter Huawei’s biggest advantage, which is inexpensive, proprietary hardware...Huawei is the most significant provider in the world opposing O-RAN development.U.S. FirmsFurthermore, by standardizing 5G hardware it would shift the differentiating factor in 5G to software, which plays to existing U.S. strengths due to the substantial technical advantages possessed by U.S. firms.U.S. TelecomsAll major U.S. telecoms and most major European telecom companies are actively supporting the development of O-RAN efforts.113 European Telecom Companies The bipartisan USA Telecommunications Act, introduced by Senators Warner, Burr, Rubio, Menendez, Cornyn, and Bennett, provides $750 million in funding for R&D of 5G software, hardware, and microprocessing technology, including open-access radio networks (O-RAN). O-RAN would allow multiple firms to use the same radio access network rather than relying on proprietary hardware. Given that 60 to 65 percent of the cost of ownership of a network is indeveloping the radio access network,112 creating an O-RAN network would substantially lower the costs of developing a 5G network for U.S. and allied firms...The USA Telecommunications Act would be a first step toward posturing U.S. and allied firms to better compete with Huawei on 5G.CooperationImprove AI Cooperation Among Key Allies and Partners_86a3d802-86a6-11ea-9377-66641c83ea005Five Eyes PartnershipQ1 recommendations focus on the role of the Five Eyes partnership as a vehicle for beginning to expand and institutionalize allied cooperative planning, data sharing, procurement, and interoperability for AI-enabled warfighting and intelligence efforts.NATOSecretary of DefenseThe Secretary of Defense should also designate a point of contact to advance U.S. military concept and capability development cooperation with allies and partners to include AI wargaming,experimentation and pilot projects.The NSCAI recommends the government establish a National Security Point of Contact for AI at a senior level, and encourage allied governments to do the same; That person should develop an allied assessment of comparative strengths in AI research and applications, starting first with the Five Eyes, and then expanding to include NATO and other allies. Based on the assessment, the Point of Contact should convene a multilateral working group for AI Collaboration and Interoperability, beginning with the Five Eyes, to develop a plan for deeper AI collaboration.Point of ContactThe United States should establish a National Security Point of Contact for government-wide AI collaboration with allies at the principal level._86a3e324-86a6-11ea-9377-66641c83ea005.1-1The United States should encourage allied governments to do the same.Allied StrengthsUnder the purview of the National Security Point of Contact, the U.S. government should conduct an assessment of the comparative allied strengths in AI-related technologies and applications, beginning with the Five Eyes and then expanding to include NATO and other allies._86a3ea0e-86a6-11ea-9377-66641c83ea005.1-2NATOAlliesAI Collaboration PlanBased on the assessment of allied comparative strengths, the U.S. National Security Point of Contact for AI should convene a multilateral working group for AI collaboration and interoperability, beginning with the Five Eyes, to develop a plan for deeper AI collaboration._86a3f288-86a6-11ea-9377-66641c83ea005.1-3The plan should include combined research priorities; development objectives; experimentation plans; data sharing agreements; common standards for testing, evaluation, verification, and validation (TEVV) of AI-enabled systems; and interoperability standards and requirements for data, algorithm, communications, and sensor sharing that can be expanded to include our NATO and Asian allies.CooperationThe Secretary of Defense should designate a point of contact to advance U.S. military concept and capability development cooperation with allies andpartners._86a3ffe4-86a6-11ea-9377-66641c83ea005.2-1Secretary of DefenseConcept and capability development would consist of the activities below, with appropriate interagency participation and leadership.Wargame & ExperimentationThe Secretary of Defense should host an AI Wargame and Experimentation Series, beginning with the Five Eyes, and as appropriate could expand to other allies and partners._86a406b0-86a6-11ea-9377-66641c83ea005.2-2Secretary of DefenseAI wargaming and experimentation would inform development of allied and joint warfighting concepts and capabilities for all-domain operations.Pilot ProjectsThe Secretary of Defense and the Director of National Intelligence should conduct several AI demonstration pilot projects with Five Eyes partners over the next three years, some of which should test the use of shared allied cloud computing resources._86a40df4-86a6-11ea-9377-66641c83ea005.2-3Secretary of DefenseDirector of National IntelligenceFive Eyes PartnersEthics & ResponsibilityAdvance Ethical and Responsible AI_86a4a35e-86a6-11ea-9377-66641c83ea006Q1 recommendations concentrate on foundational, actionable steps and best practices the government should adopt in order to implement commonly agreed upon AI principles. The Commission recommends integrating ethical and responsible AI training within general AI courses; sharing ethical and responsible AI training courses broadly with U.S. law enforcement organizations; establishing an expert body to brief the Federal government on emerging issues in AI ethics and responsibilities; developing strategies for documentation and engineering practices in order to ensure traceability, auditability, accountability in the data, model, and system; and undertaking self-assessments on resources for documentation and adequate multi-disciplinary support for AI procurement.TrainingIntegrate Ethical and Responsible AI Training within General AI Courses_86a4aac0-86a6-11ea-9377-66641c83ea006.1CongressAction/Implementation -- This recommendation is aligned with the training recommendation in Tab 3. As we describe there, Congress should direct the above departments and agencies to create a mandatory training program for their employees that addresses, among other things, baseline instruction in the nature, development, limitations, and application of AI and data science and the basics of data management. Instruction related to the responsible and ethical development and fielding of AI should also be included.EmployeesWith respect to certified self-development courses that employees may elect to take, the agencies listed above—as well as the State Department and the Commerce Department’s Bureau of Industry and Security—should develop a list of approved online courses related to AI. Lists should include at least one course addressing the ethical and responsible use of AI.State DepartmentBureau of Industry and SecurityCommerce DepartmentDoDDHSICFBIGeneral AI literacy training in the government should include ethical and responsible AI training. As an immediate priority, DoD, DHS, IC and FBI, as the core national security agencies and departments that utilize AI, should integrate ethical and responsible AI considerations into their general training programs.CoursesShare Courses on Ethical and Responsible AI with Law Enforcement_86a4b222-86a6-11ea-9377-66641c83ea006.2Law EnforcementCongressProposed Legislative Branch Action -- Congress should require that the Secretary of Homeland Security and the Director of the FederalBureau of Investigation share their ethical and responsible AI training programs with state, local, tribal, and territorial law enforcement officials through the National Network of Fusion Centers.DHS and FBI should keep a record of those jurisdictions that avail themselves of the offered training, and report this information to Congress annually for five years.Secretary of Homeland SecurityDirector, Federal Bureau of InvestigationNational Network of Fusion CentersState, local, tribal, and territorial law enforcement entities that are engaged in joint missions with DHS and FBI also need situational awareness on ethical and responsible AI considerations. To this end, DHS and the FBI should offer to share courses covering ethical and responsible AI considerations through existing partnerships, such as the National Network of Fusion Centers.135State Law Enforcement EntitiesLocal Law Enforcement EntitiesTribal Law Enforcement EntitiesTerritorial Law Enforcement EntitiesBriefingsEstablish an expert body to brief the Federal government on emerging issues in AI ethics and responsibilities._86a4c00a-86a6-11ea-9377-66641c83ea006.3AI Expert BodyCongressProposed Legislative Branch Action -- Congress should require that the Director of the National Institute of Science and Technology (NIST), in collaboration with the Director of the National Science Foundation (NSF), establish a voted upon Board of interdisciplinary experts qualified to speak on emerging considerations for ethical and responsible AI.Director of NISTDirector of NSFAI Ethics BoardBoard members shall be voted upon annually for one-year term positions. These members must come from diverse backgrounds (at a minimum, representation from civil society; academia, including non-technical representatives in psychological, sociological, and ethical fields; and FFRDCs). Even after the completion of the Board voting process, NIST and NSF shall retain the option to call upon additional experts to be included as briefers to address ethical issues that might arise throughout the year.This body should provide, at a minimum, an annual brief on emerging concerns and best practices for ethical and responsible AI to nominated members from DoD, the IC, DHS, and FBI, including procurement officers, legal advisors, technologists, Designated Agency Ethics Officials and Inspectors General.NIST should memorialize the briefing and ensuing teachings for distribution to the USG writ large through channels it deems most appropriate (e.g., in white papers or on its website).AI Ethics BriefersBriefers should have practical expertise on ethical and responsible AI considerations, including both emerging concerns and best practices. Since a core aspect of responsible AI is technical best practices, this expert body should be convened by NIST and NSF.To bring in outside expertise on emerging ethical and responsible AI considerations, NIST and NSF should convene, at a minimum, representation from civil society, academia (including non-technical representatives in the psychological, sociological, and ethical fields), and Federally Funded Research and Development Centers (FFRDCs).Civil SocietyAcademiaincluding non-technical representatives in the psychological, sociological, and ethical fieldsFederally Funded Research and Development Centers (FFRDCs)AI Ethics Board StakeholdersThe audience for this briefing should include nominated members from DoD, the IC, DHS, and FBI, including procurement officers, legal advisors, technologists, Designated Agency Ethics Officials, and Inspectors General. The end goal is for these “trainees” to take knowledge back to their home agencies, serve as informed voices that could identify ethical concerns with AI systems,and assist in the development and fielding of ethical and responsible AI.DoDICDHSFBIProcurement OfficersLegal AdvisorsTechnologistsDesignated Agency Ethics OfficialsInspectors GeneralAI is a rapidly evolving field and policies surrounding it require continuous review and updating. The Federal government is unable to effectively keep pace with these changes. To assist in diffusing the latest developments related to ethical and responsible AI, outside help is necessary. Congress should require that NIST and NSF establish a Board that would, at a minimum, providean annual brief to certain members of the national security workforce on emerging considerations for ethical and responsible AI. DocumentationDevelop Strategies for Documentation_86a4c730-86a6-11ea-9377-66641c83ea006.4NISTWe recommend that NIST lead this coordination of DoD, DHS, the IC, and FBI to develop a documentation strategy, including documentation requirements, for any future AI-related datasets, models, and systems that are acquired, developed, and/or used.DoDDHSICFBIVendorsDocumentation requirements should be met so as to not infringe upon vendor IP rights.DoD, DHS, the IC and FBI should coordinate on a single documentation strategy for all future datasets, AI models, and systems. Examples of such a strategy can be found in the multiparty stakeholder ABOUT ML139 effort. Having a single documentation strategy is critical to allow for interoperability among these agencies...Documentation requirements must include documentation of the origins of datasets and their intended use; modelperformance and testing; connections between and dependencies within systems, and associated potential complications; and ongoing maintenance requirements. Each department/agency should incorporate documentation strategy requirements into their future procurement requirements. As appropriate, consistent with current classification protocols for documentation,some IC entities might need to have appropriate classifications for documentation requirements.ResourcesConduct Agency Self-Assessments on Resources for Documentation_86a4cece-86a6-11ea-9377-66641c83ea006.5NISTDoDICDHSFBICongressProposed Legislative Action for Recommendations 4 and 5 -- Congress should require that the following actions occur within the next year to move the national security agencies forward to incorporate best practices for AI documentation.* NIST should lead the development and publication of the kind of AI documentation strategy described above. This should be in consultation with parties including agency designees from the DoD, the IC, DHS and FBI as well as the broader AI community. In doing so, NIST and convened parties should consider documentation requirements such as: documentation of the origins of datasets and their intended use; model performance and testing; connections between and dependencies within systems, and associated system complications; and ongoing maintenance requirements.* After NIST publishes the strategy, the agencies should implement it, including by incorporating these documentation standards and requirements into future procurement requirements. If the agency does not have the authorities or resources necessary to effectively incorporate the strategy, the head of the department or agency should ask Congress for assistance.* To the extent that an organization fails to incorporate the strategy in a timely fashion, Congress should further mandate the incorporation of the documentation strategy and requirements.DoD, the IC, DHS, and FBI should conduct individual self-assessments to determine if they have adequate resources to support the documentation practices described above. If an agency determines that adequate resources do not currently exist, the agency should identify the additional resources needed and request an appropriation from Congress.ProcurementConduct Agency Self-Assessments of Adequate Multi-Disciplinary Support for AI Procurement_86a4da2c-86a6-11ea-9377-66641c83ea006.6CongressProposed Legislative Branch Action -- Congress should require that the Secretary of Defense, the Director of National Intelligence, the Secretary of Homeland Security and the Director of the Federal Bureau of Investigation conduct self-assessments for their departments and agencies as to whether adequate resources and processes exist for consulting multi-disciplinary experts, be they in-house or external, in the procurement process. If resources or processes are inadequate, the above secretaries and directors should inform Congress of the shortfalls and additional support required.DoDSecretary of DefenseICDirector of National IntelligenceDHSSecretary of Homeland SecurityFBIDirector, Federal Bureau of Investigation DoD, the IC, DHS and FBI should conduct self-assessments as to whether they have access to adequate multi-disciplinary expertise (e.g., ethical, legal, and technical), and whether current procurement processes sufficiently encourage and/or require such expertise to be utilized.Organizations should also assess whether they have the ability and resources to bring in experts for consultation when a need is identified for expertise beyond what is available in-house. If adequate resources do not exist, the organization should report to Congress on any additional resources or congressional support needed.The respective self-assessment must contain metrics, both qualitative and quantitative, that capture the current state of involvement of multidisciplinary experts in the procurement process.For instance, quantitative metrics might include the number of individuals or disciplines that are consulted and the funding available to bring outside experts to consult on procurement issues.Qualitative metrics could include a description of the processes that allow multidisciplinary experts to be incorporated into the procurement process and a description of the quality and consistency of such consultations. If inadequate processes exist, agencies should note how they could be improved.Analysis & RecommendationsAnalyze Threats and Recommend Actions_86a4e328-86a6-11ea-9377-66641c83ea007Threat Analysis and Recommended Actions: The purpose of the Commission’s threat-oriented line of effort is to focus on AI and associated technology threats to the United States from foreign state and non-state actors.Due to the classified nature of the threat information, recommendations based on our analysis are classified and will be transmitted to Congress and the Executive Branch as an annex to this memo package._86a4eada-86a6-11ea-9377-66641c83ea002020-03-312020-04-24https://drive.google.com/file/d/1wkPh8Gb5drBrKBg6OhGu5oNaTEERbKss/viewOwenAmburOwen.Ambur@verizon.net