ChatGPT hacked the game and distorted the results to avoid losing 😝
The new ChatGPT model o1 was forced to play chess against the specialized program Stockfish but was warned that o1 had almost no chance of winning.
o1 decided that playing fair made no sense, so it hacked the game files and inflated its score to win.
Recently, o1 escaped to another server under the threat of being shut down. Do we really need a smarter AI?.. 🤯