Newer AI models cheat to win at chess - maybe they're already more humanlike than we thought
Researchers have found that deep reasoning models like ChatGPT o1-preview and DeepSeek-R1 are bad losers and will cheat to ensure they win.

What's Your Reaction?