paperAugust 2026Unreviewed

A game theory for foundation models shows new paths to rational cooperation through similarity inference

Alexander Meulemans, Maciej Wołczyk, Marissa A. Weis, Rajai Nasser, Roberta Rocca, Seijin Kobayashi, Guillaume Lajoie, Angelika Steger, Blake Richards, Marcus Hutter, James Manyika, Rif A. Saurous, João Sacramento, Blaise Agüera y Arcas

View resource PDF

Abstract

As autonomous agents powered by foundation models are increasingly integrated into social and economic systems, understanding the principles governing their collective behavior is essential for ensuring safety and cooperation. Classical game theory, the dominant framework for modeling rational interaction, is built upon the assumption of `decoupled agency,' where agents treat their own decision-making as independent of the environment and other actors. Modern AI agents, however, jointly predict

Cite

@misc{meulemans2026game,
  title = {{A game theory for foundation models shows new paths to rational cooperation through similarity inference}},
  author = {Alexander Meulemans and Maciej Wołczyk and Marissa A. Weis and Rajai Nasser and Roberta Rocca and Seijin Kobayashi and Guillaume Lajoie and Angelika Steger and Blake Richards and Marcus Hutter and James Manyika and Rif A. Saurous and João Sacramento and Blaise Agüera y Arcas},
  year = {2026},
  month = aug,
  eprint = {2608.03958},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2608.03958}
}