Public AI Seminar
We organize a regular research seminar with two objectives: To share research, ideas, and tools that support the development of public AI, and to foster a community capable of describing, guiding, and implementing public AI.
The first half of each session, reserved for invited talks, is recorded via Zoom and posted online. Notes are summarized without attribution, under Chatham House rules. Attendance is invite-only to keep it cozy; to join a future season, you can apply here. You can also suggest future speakers you would like to see.
Meetings are Tuesdays from 11am – 1pm ET over Zoom.
Recordings and notes are shared publicly on this page.
Season 3
Season 3 ran from April 22 to June 10, 2025, with two special sessions later in the summer. The seminar surveyed positions on AI emerging from art and culture, and considered the relationship between AI and the creative industries through four lenses: political power, cultural power, technological power, and economic power. Several spots were reserved for creative professionals in any medium interested in exploring public AI.
Case studies of Public AI are denoted by a ⭐
Date | Seminar | Recording |
---|---|---|
Apr 21 | Introductions & goals of the seminar | |
Apr 22 | Do Not Train, with Mat Dryhurst (artist), and AI and copyright, with Reema Selhi (DACS, British Copyright Council) | Video, Text |
Apr 29 | Generative AI is not useful for making art, with Ted Chiang (author) | |
May 6 | ⭐ Community data and the case of Common Voice, with EM Lewis-Jong (Mozilla) | Text |
May 13 | A Vision for Public AI in California, with Teri Olle and Taylor Jo Isenberg (Economic Security Project) | Video, Text |
May 20 | ⭐ Public interest AI, with Martin Tisné (AI Collaborative, Current AI) | Video, Text |
May 27 | AI & entertainment industry, with David White (3CG Ventures), and AI & the freelancer economy, with Angie Kim and Jessica Mele (CCI) | Video, Text |
Jun 3 | Past and future narratives for tech, with Gideon Lichfield (journalist) and Nils Gilman (historian) | Video, Text |
Jun 10 | The near possible future of AI, with Kim Stanley Robinson (author) | Video, Text |
Jul 11 | ⭐ The National Deep Inference Fabric, with David Bau (Northeastern) | Video, Text |
Aug 21 | Public strategies for AI Agents, with Dan Zhao | Vid, Memo |
Organizers: Joshua Tan (Metagov), Alek Tarkowski (Open Future), B Cavello (Aspen Digital)
Season 2
Season 2 ran from August 13 to October 15, 2024. This season focused especially on connecting and relating a cluster of political narratives around public AI. Narratives included those centered around responsible/safe AI, public compute, sovereign/national AI, democratic AI, open source AI, and AI for science.
Date | Seminar | Recording |
---|---|---|
Aug 6 | Introductions & goals of the seminar | |
Aug 13 | ⭐ Towards a network of publicly-funded AI labs, with Yoshua Bengio (MILA) | Video, Text |
Aug 20 | ⭐ Public compute, with Nicole DeCario (Allen Institute, OLMo) and Katie Antypas (National Science Foundation, NAIRR) | Video, Text |
Aug 27 | ⭐ SEA-LION, with Leslie Teo (AI Singapore) | Video, Text |
Sep 3 | AI and the labor market, with Julia Lane (NYU) and Adam Leonard (Texas Workforce Commission) | Video, Text |
Sep 10 | The role of openness in AI, with Irene Solaiman (HuggingFace) | Video, Text |
Sep 17 | Democratic AI, with Divya Siddarth (Collective Intelligence Project) | Video, Text |
Sep 24 | ⭐ GPT-SW3, with Magnus Sahlgren (AI Sweden) | Video, Text |
Oct 1 | Community power in AI, with Jeni Tennison (Connected by Data) | Video, Text |
Oct 8 | The AI Dilemma, with Aza Raskin (Center for Humane Technology) | Video, Text |
Oct 15 | AI Nationalisms, with Sarah Myers West and Amba Kak (AI Now) | Video, Text |
Organizers: Joshua Tan (Oxford, Metagov), Alex Krasodomski (Chatham House), B Cavello (Aspen Digital)
Season 1
The first season of the seminar ran from January to March, 2024. We considered different models of public AI, lessons from both history and from AI, general theories of public investment and of regulation, as well as many different arguments for and against public AI. Four themes guided our considerations: public benefit, public accountability, market failure, and political narratives.
Date | Seminar | Recording |
---|---|---|
Jan 9 | Public options for AI, with Bruce Schneier (Harvard) | Video, Text |
Jan 16 | Public policy for tech, with Diane Coyle (Cambridge) | Video, Text |
Jan 23 | ⭐ Case study: AuroraGPT & BritGPT, with Rick Stevens (Chicago / Argonne National Labs) and Hannah O’Rourke (Labour Longterm) | Video, Text |
Feb 6 | The politics of AI, with Julia Angwin (New York Times / Harvard) | Video, Text |
Feb 13 | Industrial organization and antimonopoly, with Ganesh Sitaraman (Vanderbilt) | Video, Text |
Feb 20 | Lessons from digital public infrastructure, with David Eaves (UCL) | Video, Text |
Feb 27 | AI and democracy, with Lawrence Lessig (Harvard) | Video, Text |
Mar 5 | Synthesis, conclusions, where we go from here |
Organizers: Joshua Tan (Oxford, Metagov), SJ Klein (Berkman, Concordance), Nick Garcia (Public Knowledge)