Center for Human-Compatible AI

CHAI’s workshop site announces the dates and describes the 10th annual workshop as a Chatham House Rule retreat to examine emerging AI safety research questions.

CHAI researchers introduce “Learning to Yield and Request Control (YRC),” including an open-source benchmark, a validation approach, and comparisons of learning methods across diverse environments.

CHAI affiliate Brian Christian publishes an article in Daedalus on how AI alignment has progressed toward “care-like relationships,” co-edited under “The Social Science of Caregiving.”

CHAI senior scientist Jonathan Stray proposes a practical definition of AI neutrality (“maximum equal approval”) and describes related evaluation ideas and concerns about political influence.

A CHAI update on a paper (presented at a NeurIPS 2024 workshop) studying whether allowing an agent to ask for help can mitigate goal misgeneralization under distribution shift.

A CHAI update on a paper published at the NeurIPS SoLaR workshop describing a methodology to reduce sycophancy-like unwanted LLM behaviors.

CHAI reports that Rachel Freedman, a CHAI PhD student, was selected for Cooperative AI’s inaugural PhD Fellow Program.

CHAI highlights a NeurIPS 2024 Pluralistic Alignment Workshop-accepted paper proposing a “representative social choice” framework for aligning AI with diverse human preferences.