Stanford University

Stanford Trustworthy Artificial Intelligence Research (STAIR) works to develop the principles and practice of trustworthy artificial intelligence (AI) to inform public and policy discourse on the opportunities, risks, and impacts of AI. The award supports the AI Measurement Convening, which addresses the critical gap between technical AI performance claims and real-world capabilities by bringing together computer scientists, psychometricians, ethicists, and civil society advocates to develop a shared framework that will enable informed public discourse and support evidence-based AI governance.

The Edward Ginzton Laboratory at Stanford University is an interdisciplinary research lab that explores the most dynamic and fruitful areas at the intersection between engineering and the sciences.

This project is an experiment in making and teaching at the intersection of traditional craft, digital systems, and physical modeling and leads to a public exhibition and symposium at Stanford University.  Mabuchi and Utterback are also documenting and sharing their work via an online blog, and they plan to publish a paper in the art+science journal Leonardo.  Their approach is informed by Camille’s expertise in creating interactive systems that use algorithms and Hideo’s expertise as both a physicist and a ceramicist and weaver, with the aim of amplifying the value of working with material things in the formulation of digital art workflows. The two MacArthur Fellows plan to develop and co-teach a new undergraduate course based on enmeshing digital workflows with hand-weaving.  A public symposium is planned to include additional speakers in the summer or fall of 2026. 

Stanford Trustworthy Artificial Intelligence Research (STAIR) works to develop the principles and practice of trustworthy artificial intelligence (AI) to inform public and policy discourse on the opportunities, risks, and impacts of AI. The award supports the launch of STAIR’s Evaluation and AI Governance project, which seeks to develop rigorous approaches for studying generative AI models from a public interest perspective to inform policymaking, practice, and public understanding of AI. The work will include research that seeks to elicit the values and biases of generative AI models, understand the limitations of generative AI models, and establish guidelines for generative AI model evaluations that reflect public interest considerations. The Evaluation and AI Governance project will produce a mix of academic papers, guidance and analysis for policymakers, and public commentary based on the research that is undertaken.

Stanford University is the hosting body for this collaboration between physicist and Stanford Professor of Applied Physics Hideo Mabuchi (2000 MacArthur Fellow) and sculptor and installation artist Judy Pfaff (2004 MacArthur Fellow).
This collaborative project, Colorspace , is based primarily on the genial will to work together across drastic incongruities in perceptual focus and creative habit.  Mabuchi and Pfaff begin their work together in the studio focusing on ceramic glazes, developing polychromatic formulations that can liquify and run together in the high heat of the kiln.  Pfaff 

often employs two-part liquid foams to create miscible streams of oozing color, but the fired melt and flow of ceramic glazes bring a distinct materiality to her work and pose unfamiliar challenges in process and scale.  Mabuchi is well-steeped in the technical side of ceramics but has not engaged deeply with applied color in his prior creative work.  Mabuchi is also interested in non-scientific poetics of color, as have been developed by humanists such as Goethe and Bachelard; his practice in microphotography of ceramic surfaces can be a rich source of visual cues for the Colorspace project.  This collaboration with Pfaff provides a unique opportunity for studio work with an exceptionally broad base of inspiration and two very distinct sets of “filters” for recognizing the inception of interesting new ideas.
This interaction culminates in a public exhibition in the fall of 2022, to be staged in Tivoli, NY.   Mabuchi plans to narrate his observations of the collaborative process  in an ongoing weblog openly available to the public. The public exhibition is expected to draw residents and visitors from the New York City area and the greater Hudson River Valley.  

The Stanford Cyber Initiative is an interdisciplinary academic center, drawing on Stanford University’s strengths in computer science, law, public policy, and other relevant disciplines. The award supports the creation of an Intergovernmental Personnel Act mobility program fellow within the Cyber Initiative that will provide guidance within the federal government on three key policy areas related to free expression and privacy: 1) federal debates about whether the government should mandate access to plaintext versions of encrypted information; 2) ongoing reforms related to U.S. government surveillance activities; and, 3) reforming federal policies that are hostile towards digital security research.