)
Initiative: Human-Centered Artificial Intelligence
A leap for humankind
When human beings landed on the moon, a seemingly impossible feat was made possible by uniting the brightest minds across various disciplines. This moment in AI holds similar promise to expand the horizons of human potential. To realize this potential, we must weave human-centered principles into every facet of AI development, making ethical, equitable, and impactful AI the standard, not the exception.
The Stanford Institute for Human-Centered Artificial Intelligence (Stanford HAI) is an interdisciplinary initiative dedicated to advancing AI research, education, policy, and practice to improve the human condition. Drawing on expertise from faculty across the university, including Stanford’s renowned computer science department, and collaborating closely with industry and government, the institute aims to guide the development of AI systems that are groundbreaking, ethical, and equitable—an AI moonshot to serve and uplift humanity.
)
United we innovate
Stanford HAI’s leaders share a commitment to the mission of guiding human-centered AI.
)
Fei-Fei LiStanford HAI Denning Co-Director)
James LandayStanford HAI Denning Co-Director)
John EtchemendyStanford HAI Denning Co-Director
REAL-WORLD SOLUTIONS
Stanford HAI is uniquely positioned to leverage Stanford’s world-class faculty, collaborative culture, and depth of knowledge across fields as diverse as computer science, engineering, human health, and public policy. We tackle the most pressing challenges our world faces today by investing in faculty and students, research grants, and policy work.
)
A moonshot moment
Imagine a world where AI helps eradicate hospital-acquired infections, tailors education to individual student needs, and detects forced labor at remote worksites. At Stanford HAI, we are turning these possibilities into reality. We are reaching for the stars in AI innovation, harnessing the brilliance of interdisciplinary teams to revolutionize technology for the good of humanity.
)
Daniel Ho
Influencing public policy
Policymakers are crucial for AI’s success but often lag behind industry and academia when it comes to understanding this technology. At Stanford HAI, we actively engage with state, national, and global leaders to shape policies that mitigate AI’s risks and maximize its benefits. From serving as the State of California’s chief AI partner to advising the United States Congress and participating in United Nations panels, our faculty influence policymaking at every level. In addition to advising policymakers, we have trained more than 80 congressional staffers through our bipartisan congressional boot camps and more than 8,000 government employees through symposia and other educational programs.
Auditing the IRS for racial bias
Daniel Ho, a senior fellow at Stanford HAI and the Wm. Benjamin Scott and Luna M. Scott Professor of Law, uncovered that the IRS is three to five times more likely to audit Black taxpayers due to flaws in its algorithmic audit selection process. This revelation sparked a congressional outcry and a commitment from the IRS to remedy the discrimination. Ho and his team developed predictive models to identify potential tax evasion more accurately, reducing unnecessary audits and concentrating resources on high-risk cases. “Researchers did a great service spotting the racial bias in the algorithms that guide audit selection,” said Senate Finance Committee Chairman Ron Wyden.
)
Extinguishing modern slavery
Funded by an HAI seed grant, Stanford’s Human-Trafficking Data Lab developed an AI system that spots forced labor at deforestation sites in Brazil’s Amazon rainforest. Both legal and illegal deforestation efforts routinely rely on forced labor, but the sites are remote and transitory, often abandoned before inspectors can reach them. The lab’s system addresses that challenge by recognizing one of the telltale signs of an active deforestation site: large ovens that convert the fallen timber into charcoal. The researchers are working with the Brazilian Federal Labor Prosecution Office to develop a system that automatically notifies investigators, enabling rescuers to mobilize more rapidly.
)
Improving health
As society grapples with an aging population and a shortage of health care labor, AI has the potential to revolutionize medical diagnostics, treatment, and patient care. However, it also raises critical ethical and privacy concerns. Drawing on expertise from medicine, computer science, ethics, and policy, Stanford HAI aims to develop AI that enhances health care quality for all patients. The RAISE Health initiative, a partnership between Stanford Medicine and HAI, promotes ethical AI use in health care by fostering collaboration, sharing resources, and establishing standards for safe, equitable practices.
)
An X-ray is worth a thousand words
During the COVID-19 pandemic, the Stanford Center for Artificial Intelligence in Medicine and Imaging (AIMI), led by HAI Senior Fellow Curt Langlotz, developed AI models to detect COVID-19 from chest X-rays and CT scans. These models were critical in managing the pandemic, particularly in resource-limited areas, by enabling rapid and accurate diagnosis, predicting disease progression, and assessing patient outcomes. Langlotz’s frequently quoted witticism drives home AI’s game-changing potential in this space: “‘Will AI replace radiologists?’ is the wrong question. The right answer is: ‘Radiologists who use AI will replace radiologists who don’t.’ ”
Robots that help you find your footing
Falls are the leading cause of injury among adults over age 65, with about 36 million incidents reported each year in the United States, often resulting in devastating consequences. To combat this, Stanford HAI researchers are developing an AI-powered wearable exoskeleton designed to predict and prevent falls before they occur. Led by C. Karen Liu, a computer science professor and HAI affiliate, a multidisciplinary team is creating a robotic system that helps users maintain their balance in real time—from adjusting a step to avoid a stumble to detecting hazards like uneven ground. Supported by a Hoffman-Yee Grant, this project aims to turn AI into a proactive helping hand for at-risk adults, fostering safer mobility and greater independence.
)
Professor Liu directs The Movement Lab, where she’s working to build predictive human motion models for preventing musculoskeletal injury.
)
HAI AI4ALL summer camp
Customizing education
Stanford HAI equips students and professionals with the knowledge to develop and apply AI responsibly through interdisciplinary courses, workshops, and seminars integrating computer science with the humanities, social sciences, and ethics. In partnership with the Stanford Graduate School of Education and the Stanford Accelerator for Learning, HAI hosts the annual AI+Education Summit, uniting educators, industry executives, and researchers to explore and advance the role of AI in transforming education. HAI also prepares undergraduates to become future leaders in AI through a wide range of courses, such as AI Awakening with Erik Brynjolfsson and Human-Centered AI with Peter Norvig.
Math made personal
Research shows that tailored instruction significantly improves students’ academic and life outcomes. However, 70 percent of educators struggle to address diverse student needs due to time constraints and inadequate tools. Dora Demszky, an assistant professor in education data science, is addressing this by combining machine learning, natural language processing, linguistics, and practitioner input to create scalable, equitable, and student-centered educational tools. Supported by a Stanford HAI seed grant, Demszky partnered with a network of school districts to develop AI-powered resources that help middle school math teachers customize their curriculum to meet students’ specific needs.
Sending AI back to high school
“Regardless of whether someone anticipates a technical career, being an informed person in society means understanding AI,” says Victor Lee, an associate professor in the Graduate School of Education and HAI affiliate. To prepare young people for AI’s impact, Lee initiated Project CRAFT (Classroom-Ready Resources About AI for Teaching), offering free, multidisciplinary curriculum resources for high school teachers through the Project CRAFT website. The program awards co-design fellowships to high school teachers to ensure these resources will make a difference in the classroom. In addition, HAI’s AI4ALL summer camp, launched in 2022, immerses underrepresented students in hands-on AI research and mentoring at Stanford to increase diversity among the field’s future leaders. The program saw a 90 percent increase in applicants from 2023 to 2024, and the 2024 cohort was notably diverse, with four continents represented.
)
Humans in the loop
Stanford faculty bring humanistic perspectives—drawn from disciplines including literature, art, and the social sciences—to help create AI systems that are not only innovative but also ethical and equitable.
)
Ge WangHAI associate director; director of music)
Michele ElamWilliams Robertson Coe Professor in Humanities & Sciences
Scaling sustainability
Can AI’s power be harnessed to help mitigate a challenge as great as climate change? At Stanford HAI, we leverage AI to address critical sustainability issues such as energy efficiency and environmental protection with participants from HAI, the Stanford Doerr School of Sustainability, and researchers across Stanford and beyond.
)
Carbon-conscious coding
The full suite of experiments needed to build and train an off-the-shelf AI language processing system can generate 78,000 pounds of carbon emissions. In collaboration with Facebook AI Research, Stanford HAI-affiliated faculty created an easy-to-use tool to measure AI projects’ electricity use and carbon emissions, helping developers steer toward greener algorithms.
Looking before we leap into mining projects
As clean energy transitions accelerate, the need for
critical minerals for battery metals—sourced primarily from rainforests—raises environmental and human concerns. Initially funded by a Stanford HAI seed grant, the Mineral-X Initiative enables companies to assess water, energy, and land uses to identify unsuitable projects in the early exploration stages rather than at the later mining stages. The initiative also works to safely and efficiently capture and store carbon emissions.
)
)
Erik Brynjolfsson
Reshaping the economy
Customer service chatbots, supply chain optimizing robots, and rideshare apps—AI is already transforming the economy, and its influence is set to grow exponentially. Led by HAI Senior Fellow and Jerry Yang and Akiko Yamazaki Professor Erik Brynjolfsson, Stanford HAI’s Digital Economy Lab brings together researchers and experts to explore how digital technologies are reshaping work, organizations, and the economy. By analyzing the impacts of AI, automation, and digital platforms on labor markets, productivity, and economic growth, the lab provides critical insights to inform policymakers, businesses, and the public on harnessing digital innovations for inclusive and sustainable economic development.
A helping hand, not a pink slip
How can we ensure that AI tools will augment rather than replace human workers? To inform AI implementation strategies in the workplace, the Digital Economy Lab studied the introduction of a generative AI conversational assistant using data from 5,179 customer support agents. The findings revealed a 14 percent increase in productivity on average, with the greatest benefits seen among novice and low-skilled workers due to the dissemination of knowledge from experienced colleagues. In addition, AI enhanced customer sentiment, reduced the need for managerial intervention, and improved employee retention, highlighting AI’s potential to transform the workplace positively.
The cost of misinformation
When AI is used to amplify misinformation, the consequences can be severe—from exacerbating public health crises to inciting political violence. The Digital Economy Lab investigated the financial mechanisms fueling the spread of online misinformation, focusing on how advertising revenue sustains misinformation outlets. The research uncovered that many companies inadvertently advertise on these outlets due to automated ad distribution. Furthermore, public awareness of a company’s unintentional role in funding misinformation can decrease demand for its products. The lab’s proposed low-cost, scalable interventions for digital platforms aim to curb the monetization of misinformation and reduce its spread online.
Stories:Shape the future
Whether reducing racial bias in algorithmic systems, advancing sustainable practices, or curbing the spread of misinformation, Stanford HAI strives to create solutions that benefit all.
)
Ambient intelligence, human impact
Health care providers struggle to catch early signals of cognitive decline. AI & computational neuroscientist Ehsan Adeli’s innovative computer vision tools may offer a solution.
Story tags:
:focal(513x175:514x176))
Artful intelligence
How can we develop artificial intelligence that respects our humanity?
Story tags:
:focal(3629x1117:3630x1118))
Using machine learning to help refugees succeed
How GeoMatch is revolutionizing resettlement efforts.
Story tags:
)
This)
isthe moment
for Human-Centered Artificial Intelligence
Aligning AI’s remarkable advances with human values isn’t just a goal for us—it’s a moral imperative. Our success isn’t measured by what happens here alone: We aim to ensure that people worldwide benefit from AI. Whether reducing racial bias in algorithmic systems, advancing sustainable practices, or curbing the spread of misinformation, Stanford HAI strives to create solutions that benefit all.
)
Initiatives:Explore more
)
Discover initiatives
The scope of Stanford’s vision matches the scope of the challenges we face. Because Stanford has built excellence in an unusually broad range of fields, and because we’ve connected those fields.