Stanford AI Scholars Find Support for Innovation in a Time of Uncertainty

For decades the United States has been a leader in science and medicine, largely due to robust federal funding. Today that vital support is weakening. Cuts to funding from organizations including the National Institutes of Health (NIH) and the National Science Foundation (NSF) have stalled valuable scientific and medical research programs across the country. These cuts will slow advances in new medicines and transformational technologies and risk diminishing the U.S.’s appeal as a destination for the world’s top scientific talent.

To support Stanford faculty and students during this funding challenge, the Stanford Institute for Human-Centered AI (HAI) continues to allocate substantial resources for groundbreaking projects and emerging ideas in the space. Now in its sixth year of operations, the institute draws funding from a wide variety of individual donors, philanthropic organizations, and industry partners, in addition to government grants.

“Partnerships between academic institutions, government, and industry have always been essential to driving innovation that benefits society,” says Vanessa Parli, director of research for HAI. “Many researchers are losing the resources they need to investigate ambitious ideas and test new hypotheses, making HAI’s grant programs more important than ever.”

Since its founding in 2018, HAI has distributed $50 million to more than 400 faculty across all seven Stanford schools. The institute’s various grant and fellowship programs are designed to equip Stanford faculty with the resources they need to ensure a bright future for AI innovation. Here is an overview of the various funding mechanisms HAI makes available:

Hoffman-Yee Research Grants

HAI’s largest funding resource, the Hoffman-Yee Research Grants funded by philanthropists Reid Hoffman and Michelle Yee, helps to launch “asteroid shot” ideas that address scientific, technical, or societal challenges requiring an interdisciplinary team and a bold approach. Applicants must demonstrate how their work ties into HAI’s key research pillars of understanding the human and societal impact of AI, augmenting human capabilities, and developing AI technologies inspired by human intelligence. Each winning team receives up to $500,000 in year one, with the opportunity to receive up to $2 million more over the following two years.

In 2022, a group of scholars led by HAI Senior Fellow Michael Bernstein, an associate professor of computer science, proposed to explore the possibility of encoding societal values into social media algorithms without sacrificing the core of what can make social media compelling. “The Hoffman-Yee funding let us hit the ground running and demonstrate the results that launched the entire effort,” he says. The team has since developed techniques for embedding societal values into social media feed ranking algorithms and deployed the intervention to over 1,000 people in a field experiment where they re-ranked individuals’ Twitter (X) feeds in real time. Next up, the team is creating tools to allow end users to shape their own social media experience.

“The ability to deploy our tools at scale is a rare opportunity for scientific research,” Bernstein says. Moreover, the grant has created a durable, tight set of collaborations across Stanford departments. “HAI's value of prioritizing projects that span schools and departments is a major driver of our advances. We have multiple students, postdocs, and faculty working on each of these projects in collaborations that would be difficult to sustain in any other way.”

Seed Research Grants

The HAI Seed Research Grant program supports researchers in the earliest stages of exploring speculative AI ideas across a wide range of subjects and fields. Projects can take the form of discrete studies, book-length research initiatives, a speaker series, or system building and evaluation. The goal of this funding is to help teams get initial results from their research. 

According to HAI Faculty Affiliate Jonathan H. Chen, an assistant professor of medicine, “Seed research grants from HAI have been critical to supporting our work at the HealthRex Lab. Federal funding is important, but it’s not the only way to succeed in biomedical informatics research.” For one project, Chen and colleagues received an HAI and AIMI Partnership Grant to study digital machine learning prediction models for oncology diagnostic testing that could help oncologists make the best choice for their patients, and this work has recently been accepted for publication in Nature Digital Medicine. “The interdisciplinary collaboration that HAI facilitates has been essential to our ability to learn from other domains,” he adds.

Graduate and Postdoctoral Fellowships

In addition to its grant programs, HAI supports promising researchers whose work cuts across different fields and risks being overlooked by rigidly defined academic departments. The Postdoctoral Fellowship Program offers opportunities to explore topics, conduct research, and collaborate across disciplines related to AI technologies, applications, or impact. Current HAI Postdoctoral Fellow Joba Adisa, for instance, has teamed with Victor Lee, an associate professor at the School of Education, to promote AI literacy and offer free resources to high school teachers through the CRAFT project.

“HAI provides a built-in network to bring your ideas to life. It’s a diverse community that includes mentorship, events, and opportunities to collaborate with amazing people as we explore how to frame AI in ways that augment human capabilities,” Adisa says.

HAI also offers Stanford graduate students a three-quarter fellowship designed to encourage interdisciplinary research. The program fosters collaboration between engineers, social scientists, and others researching the future of purposeful, intentional, and human-centered AI.

Faculty Fellowships

HAI’s first faculty fellow, Johannes Eichstaedt, studies the intersection of psychology and AI through a full-time, five-year appointment, and he’s been instrumental in positioning Stanford as a leader in examining AI for mental health. Reflecting on the resources and benefits of this appointment, he says, “The Stanford ecosystem is unparalleled in supporting its faculty. The university is full of stellar colleagues who push and challenge one’s research. And with startup funds, seed grants, and cloud credits, HAI and Stanford have been wonderfully generous in helping to incubate my interdisciplinary research.” Eichstaedt is especially grateful to HAI for connecting him with policy, industry, and science partners to think through how AI is changing society socially and psychologically— the focus of the book he’s currently writing.

A second HAI faculty fellow, Hari Subramonyam, an assistant professor of education, joined the institute in 2021 to advance work that pulls from human-computer interaction and the learning sciences. He focuses on using AI to augment human learning, creativity, and sensemaking by incorporating principles from cognitive psychology.

HAI also appoints faculty fellows who are jointly supported by the institute and one of Stanford’s seven schools to further the goal of multidisciplinary research. For example, Erik Brynjolfsson, director of the Stanford Digital Economy Lab, is funded by HAI and the Graduate School of Business. Yejin Choi, a former senior director at NVIDIA, was appointed HAI senior fellow in January, with additional support from the School of Engineering.

Google Cloud Credit Grants

Through a partnership with Google, HAI accelerates innovative ideas that require advanced computational resources from the commercial cloud. These Cloud Credit Grants offer Stanford researchers the chance to access up to $100,000 of Google Cloud credits— generally more than the company allocates through its own Google Higher Education programs.

 One recent project used cloud credits to create a new social reasoning benchmark for large language models. Another team received support for a novel approach to training discrete diffusion models, which outperformed GPT-2 in testing and won best paper at the International Conference on Machine Learning (ICML) 2024.

HAI Workshops

HAI is also able to support faculty through workshops that convene scholars and other experts to delve into areas of shared interest. With help from the institute, Judith Fan, an assistant professor of psychology, will soon host a workshop on advancing the science of data and visualization literacy. Dubbed the Nightingale Workshop, the two-day event will gather researchers, technologists, and educators to explore an issue that is vital for enhancing public understanding of science. “HAI has been incredibly supportive of our community-building efforts. The institute has stepped in to help, just as future support from the NSF is increasingly uncertain.” 

Recent HAI workshops have covered subjects as varied as Interactive AI Systems for Live Audiovisual Performance and The First Workshop of a Public AI Assistant to World Wide Knowledge. Check Upcoming Events for future HAI workshops.

Student Affinity Groups

To help build the next generation of AI scholars, HAI created the Student Affinity Groups program for cross-disciplinary teams to probe human-centered AI themes of their choice. Funding for basic expenses is available for any group of Stanford students, from undergraduates to postdocs and across all seven Stanford schools, for one academic year. Robotics PhD candidate Julia Di took advantage of the program to study the future of embodied AI, while computer science postdoc Kristina Gligorić  formed a group of computer science, linguistics, and psychology scholars to examine where natural language processing intersects with computational social science.

Innovation Engine

Through its various programs, HAI has funded more than 500 projects to date, and most recipients are quick to acknowledge that the value of HAI’s support goes well beyond financial resources. By bringing teams from multiple departments together for grant submissions, HAI advances the cross-sector conversations that are essential to grappling with and ultimately solving real-world problems with AI technology.

As Eichstaedt says, “HAI’s generous funding mechanisms have become a pulse of collaboration across campus. It’s in the spaces between the disciplines where bigger impacts are more likely.”

For more examples of recent AI projects and details on fellowship and grant applications, visit HAI Research.