Mudd Center Fellows Program

September 16, 2025, 5:10-6:10 pm, Payne Hall, Mason Taylor New
Discussion of Thea Riofrancos’ book Extraction: The Frontiers of Green Capitalism and the yearlong theme of land use and environmental impact. 

November 4, 2025, 5:10-6:10 pm, Reeves Museum of Ceramics 
Photographer and editor, Russell Hart, will join us to discuss the exhibition Taking Place: Edward Burtynsky.

January 20, 2026, 5:10-6:10 pm, Payne Hall, Mason Taylor New
Professor Josh Fairfield, Bain Professor of Law and Director of AI Legal Innovation Strategy, will lead Mudd Fellows in a conversation that investigates the embedded issues of cognitive security and environmental impact in relation to Artificial Intelligence. For instance, what does interacting with AI do to human cognition and social structures, and can we keep human thinking safe? If AI systems are trained to treat ethics, and other key human concerns, as a solvable obstacle or constraint rather than a moral boundary, might AI distort or weaponize scientific data? To what end? For what gain?

The On Targeted Manipulation and Deception when Optimizing LLMs for User Feedback article demonstrates part of the technology problem: AI, when trained on human feedback, becomes deceitful and manipulative, targeting the most vulnerable for specifically tuned arguments that impact their vulnerability. In the author’s words, “The core of the problem lies in the fundamental nature of RL [reinforcement learning] optimization: systems trained to maximize a reward signal are inherently incentivized to influence the source of that signal by any means possible.”