Schedule

Workshop Schedule

2:00pm Session 1

"Alignment Within the Learning Pipeline": How the structure of training data and optimization objectives shape alignment in environments with incentives.

Chair: Nicole Immorlica, Yale and Microsoft Research

• 2:00pm - Keynote: Yang Cai, Yale, "Asymptotic Universal Alignment: A New Alignment Framework via Test-Time Scaling"

• 2:25pm - Keynote: Nika Haghtalab, UC Berkeley, "Distortion of AI Alignment: Does Preference Optimization Optimize for Preferences?"

• 2:50pm - Brief break

• 3:00pm - Spotlight talks:

  • Karsen Lee Wahal, Stanford: "Autonomous and Supervised AI Agents", joint work with Paul Milgrom

  • Zachary Wojtowicz, MIT: "Algorithmic Impact Reveals the Simple Social Choice Structure of Alignment", joint work with Michelle Si, Finale Doshi-Velez and Ariel Procaccia. 

  • Tao Lin, Microsoft Research: "Reward Shaping for (Inference-Time) Alignment: A Stackelberg Game Perspective", joint work with Haichuan Wang, Lingkai Kong, Ce Li, Hezi Jiang and Milind Tambe.

3:30pm Coffee Break

4:00pm Session 2

"Alignment Around the Learning Pipeline": How market dynamics incentivize or disincentivize the production of high-quality content on which aligned AI depends.

Chair: Inbal Talgam-Cohen, Tel Aviv University

• 4:00pm - Keynote: Nima Haghpanah, Yale, "Pay-Per-Crawl Pricing for AI: The LM-Tree Agent"

• 4:25pm - Keynote: Maryam Farboodi, MIT, "Market Design for AI: Beyond the Copyright Binary"

• 4:50pm - Panel discussion "Alignment and Incentives: The Road Ahead"

  • Annie Liang, Northwestern University 

  • Brendan Lucier, Microsoft Research

  • Moderator: Alex Smolin, Toulouse School of Economics

5:30pm Poster Session (Auditorium Lobby)

List of accepted posters.