Evaluating Prediction in Acausal Mixed-Motive Settings
This is a research note presenting a portion of the research Tim Chan completed during MATS 8.0 under the supervision of Francis Rhys Ward. TLDR 1. Acausal interactions, interactions where agents influence each other without communication or any other causal action, can be a big deal. Unaligned AIs might respect...
Aug 31, 202514
Interesting work!
It might be useful to discuss the implications a bit more.
TLDR: I think you've showed superrational propensities, which adds to existing work, but successful "collusion at a distance" and other forms of acausal cooperation[1] between different models also depends on additional, possibly quite advanced, capabilities. Without them, a superrationally-inclined LLM just expecting other LLMs to reason similarly and cooperate may get exploited by other LLMs - resulting in no mutual cooperation. There's some evidence that the post's main example of GPT-5 is vulnerable to that.
- Some recent models do tend to think about superrationality and acausal decision theories in game-theoretic setups. This seems to help them at least not defect against identical copies
... (read more)