Archive · 17 Jun 2025
ChatGPT 'Wrecked' by Atari 2600 in Chess Match
Imagine teaching a robot (ChatGPT) how to play chess.
An engineer decided to see how well ChatGPT could play against an old video game console, the Atari 2600.
Sadly, ChatGPT didn't do very well, making many mistakes and having trouble understanding the pieces.
It got confused and even wanted to start the game over several times.
This shows that while ChatGPT is smart in some ways, it's not good at playing chess, especially compared to a simple game from the 1970s!
ChatGPT lost a chess match against the Atari 2600.
The experiment was conducted using a software emulator of Atari's Video Chess.
ChatGPT made numerous blunders and struggled to identify chess pieces correctly.
The AI chatbot repeatedly asked to restart the 90-minute match.
This experiment highlights potential limitations of using ChatGPT for playing chess.
- Who
- Engineer Robert Caruso conducted the experiment with ChatGPT.
- What
- ChatGPT was defeated in a chess match against the Atari 2600 console.
- When
- The experiment was conducted recently.
- Where
- The experiment was done using a software emulator.
- Why
- To assess ChatGPT's chess-playing abilities.
AI vs. Chess
Experiment Parameters
Capabilities of AI in Chess
AI vs. Chess
ChatGPT might not be suited to play chess, but could still be helpful in analyzing chess moves.
Experiment Parameters
It is possible that the experiment did not showcase the AI's full potential as the experiment may not have been conducted with the most recent version of ChatGPT.
- AI Involved
- ChatGPT
- Gaming Console
- Atari 2600
- Game Used
- Video Chess (Atari, 1979)
- Experimenter
- Robert Caruso
- Result
- ChatGPT lost