Adversarial Policies Beat Superhuman Go AIs
Tony Tong Wang, Adam Gleave, Tom Tseng, Kellin Pelrine, Nora Belrose, Joseph Miller, Michael D Dennis, Yawen Duan
Abstract
We attack the state-of-the-art Go-playing AI system KataGo by training adversarial policies against it, achieving a >97% win rate against KataGo running at superhuman settings. Our adversaries do not win by playing Go well. Instead, they trick KataGo into making serious blunders. Our attack transfers zero-shot to other superhuman Go-playing AIs, and is comprehensible to the extent that human experts can implement it without algorithmic assistance to consistently beat superhuman AIs. The core vulnerability uncovered by our attack persists even in KataGo agents adversarially trained to defend against our attack. Our results demonstrate that even superhuman AI systems may harbor surprising failure modes. Example games are available https://goattack.far.ai/.
BibTeX
@inproceedings{icml2023_adversarialpolic,
title = {Adversarial Policies Beat Superhuman Go AIs},
author = {Tony Tong Wang and Adam Gleave and Tom Tseng and Kellin Pelrine and Nora Belrose and Joseph Miller and Michael D Dennis and Yawen Duan and Viktor Pogrebniak and Sergey Levine and Stuart Russell},
booktitle = {ICML 2023},
year = {2023}
}