About HAIOur goal is for Stanford HAI to become an interdisciplinary, global hub for AI thinkers, learners, researchers, developers, builders and users from academia, government and industry, as well as leaders and policymakers who want to understand and leverage AI’s impact and potential.Artificial Intelligence has the potential to help us realize our shared dream of a better future for all of humanity, but it will bring with it challenges and opportunities we can’t yet foresee.Stanford Institute for Human-Centered Artificial IntelligenceHAI_17cdc340-4feb-11e9-bc5f-1064861d81c0Fei-Fei LiCo-Director of Stanford Institute for Human-Centered AIJohn EtchemendyCo-Director of Stanford Institute for Human-Centered AIStanford UniversityStanford HAI leverages the university’s strength across all disciplines, including: business, economics, education, genomics, law, literature, medicine, neuroscience, philosophy and more. These complement Stanford's tradition of leadership in AI, computer science, engineering and robotics.AI ThinkersAI LearnersAI ResearchersAI DevelopersAI BuildersAI UsersAcademiaGovernmentIndustryLeadersA better future for all of humanity_17cdc502-4feb-11e9-bc5f-1064861d81c0To advance AI research, education, policy, and practice to improve the human condition._17cdc5de-4feb-11e9-bc5f-1064861d81c0ResearchAt Stanford HAI, our vision for the future is led by our commitment to studying, guiding and developing human-centered AI technologies and applications. We believe AI should be collaborative, augmentative, and enhancing to human productivity and quality of life.CollaborationAugmentationHuman ProductivityQuality of LifeImpactStudy and forecast the human and societal impact of AI._17cdc692-4feb-11e9-bc5f-1064861d81c01To develop equitable and trustworthy technology, we must understand how AI performs in practice, and guide and shape the way AI interacts with humans, their vital social structures and institutions, and the international order.Artificial intelligence and machine learning are poorly understood, even within academic and research communities. The media portray a world of robots run amok; new applications and milestones are often described as “machines beating humans;” and influential public figures warn of job losses and more dire consequences. While some concerns are legitimate, misleading narratives too often distract from the pressing issues society is likely to confront as AI systems become commonplace.Scholarly research is needed to measure and manage a host of critical issues, including the extent to which algorithms introduce, compound, or mitigate business risk or bias; a “responsibility gap” between decisions made by machines and people; the use of AI for surveillance, population control, and waging war; and the impact of AI on industry structure, labor markets, economic growth, and trade across nations. This research will inform engagement with industry, government, and civil society to beneficially guide AI’s development.Sample Research Projects:BiasLeverage machine learning to discover and correct its own biases._17cdc750-4feb-11e9-bc5f-1064861d81c01.1James ZouLonda SchiebingerSerena YeungCarlos BustamanteCorrecting Gender and Ethnic Biases in AI Algorithms -- Machine learning algorithms can contain gender and ethnic biases. As AI becomes ubiquitous, such bias if uncorrected can lead to inequities in service and discrimination against specific populations. In this project, we will develop a AI auditing where we leverage machine learning to discover and correct its own biases. Our goal is to make AI audit an integral component of machine learning in industry and academia.HumanhoodUnderstand and forecast how the increasing presence of AI in daily life will change perceptions of what it means to be human._17cdc822-4feb-11e9-bc5f-1064861d81c01.2Benoît MoninErik SantoroThe Impact of Artificial Intelligence on Perceptions of Humanhood -- Can logic remain at the core of what it means to be human if AI clearly surpasses humans at it? Will society redefine what is core to the human experience as humans lose ground to AI on cognitive abilities that traditionally enshrined humans at the top of the animal kingdom? Drawing on social psychological theory and using randomized control trial (RCT) experiments, we seek to understand and forecast how the increasing presence of AI in daily life will change perceptions of what it means to be human.CapabilitiesDesign AI applications that augment human capabilities._17cdc8e0-4feb-11e9-bc5f-1064861d81c02HAI seeks to develop new human-centered design methods and tools so that AI agents and applications are designed and created with the ability to communicate with, collaborate with, and augment people more effectively, and to make their work better and more enjoyable. These breakthroughs will allow great progress in healthcare, education, sustainability, automation, and countless other domains.AI has the potential to replace people in their jobs. But AI also has the potential to educate, train, and augment people, making them better at their tasks and activities. AI can make the quality of an individual’s work better, resulting in better writing, design, healthcare, communication, teaching, and art.People are social animals; machines are not. To achieve broad acceptance, AI systems must conform to the often-implicit cultural conventions that underlie human interaction and communication. When should such systems “listen” and when should they “speak up”? If they require a shared resource, how can they balance their own needs with those of others? If humans are asked to rely on machine guidance to augment their decisions (and perhaps override their intuition), they may need to understand the strengths and weaknesses of the AI.The advances and considerations developed in our other areas of focus, in addition to research in design methods, will help us to create systems that have these more appropriate communication capabilities. This underlying research will be combined with the use of AI in important application domains, such as education, healthcare, and sustainability, where the new design methods and tools can be leveraged and evaluated.Sample Research Projects:Facial ExpressionPrototype an AI tool for automatic facial expression recognition._17cdc9b2-4feb-11e9-bc5f-1064861d81c02.1Dennis WallTom RobinsonTerry WinogradDynamic Artificial Intelligence-Therapy for Autism on Google Glass -- Children with autism (ASD) struggle to recognize facial expressions, make eye contact, and engage in social interactions. There is potential to meet this need through wearable tools. Tapping into this potential, we have prototyped an AI tool for automatic facial expression recognition that runs on Google Glass through an Android app to deliver social emotion cues to children with autism while interacting with family members in their natural environment. With the HAI grant, we will refine the system’s efficacy and ready it for deployment.HapticsOptimize haptic guidance._17cdca7a-4feb-11e9-bc5f-1064861d81c02.2Julie WalkerAndrea ZanetteMykel KochenderferAllison OkamuraLearning Haptic Feedback for Motion Guidance -- Haptics is a promising method for providing guidance to users during human-machine interaction, particularly through wearable or ungrounded devices. We plan to apply modeling and reinforcement learning to optimize ungrounded and wearable haptic guidance. We hope these methods will improve the ability of humans and intelligent systems to communicate effectively during tasks such as robotic surgery, teleoperation, and collaborative object manipulation.IntelligenceDevelop AI technologies inspired by the versatility and depth of human intelligence._17cdcb56-4feb-11e9-bc5f-1064861d81c03Current AI systems lack flexibility and contextual understanding, and resist explanation in terms comprehensible by humans. Ultimately we need to develop machine intelligence that understands human language, emotions, intentions, behaviors, and interactions at multiple scales.Today’s AI methods can perform simple, well-defined, narrow tasks well, but only after training on laboriously annotated data. While recent algorithms have enabled us to solve formerly intractable real-world problems, it remains to be seen how far they can go, and whether they can ultimately serve as the basis for a general theory of intelligence and the development of truly intelligent machines.Current AI systems lack flexibility and contextual understanding, and resist explanation in human-comprehensible terms. To create a machine-assisted — yet human-centered — world, we must develop the next generation of AI techniques that overcomes the limitations of current algorithms, expands the class of problems that can be addressed, and complements human cognitive and analytic styles. Ultimately we need machine intelligence that leads to good decisions, either acting alone or working in combination with human decision-makers. It should understand human language, emotions, intentions, behaviors, and interactions at multiple scales.Tackling these challenges on both the theoretical and practical levels requires substantial fundamental research. Developing a next generation of human-centered machine intelligence will demand combining further research in core machine learning and artificial intelligence with approaches coming from our growing understanding of human intelligence developed in areas including neuroscience and cognitive science.Sample Research Projects:PerceptionDevelop insights into human perception._17cdcc50-4feb-11e9-bc5f-1064861d81c03.1Gregory ValiantNoah GoodmanAdversarial Examples for Humans?Humans are generally regarded as the gold-standard for robust perception and classification, and the implicit assumption in much of the work on “adversarial examples” is that humans do not have such vulnerabilities. We hope to understand whether there are settings where nearly every natural input (e.g., pertaining to vision, speech recognition, perception, etc.) can be turned into an “illusion”. We hope that this work will yield significant insights into aspects of human perception; the results may have implications for security and safety.Free ExplorationDevelop a formal methodology for reasoning about the free exploration provided by humans who interact with machine learning and AI systems._17cdcd2c-4feb-11e9-bc5f-1064861d81c03.2Mohsen BayatiRamesh JohariFree Exploration in Human-Centered AI Systems -- All systems that learn from their environment must grapple with a tradeoff between making decisions that maximize current rewards ( "exploitation") and decisions that are likely to teach the system about the environment and thus potentially increase future rewards ("exploration"). Automated machine learning and AI systems leverage techniques to balance exploration and exploitation to maximize rewards over time. This dynamic becomes more complicated at the interface between ML systems and humans. Our goal is to develop a formal methodology for reasoning about the free exploration provided by humans who interact with machine learning and AI systems.2019-03-26https://hai.stanford.edu/OwenAmburOwen.Ambur@verizon.net