Ever since computers came along, the processing power at their command has been frequently used for one vital purpose: playing chess. Humanity chose to engage in a war against its own creation, only to find out that the creation has toppled its own god. Yes, even dating back to the 1950s, computers have been beating humans in chess matches. But how would something like ChatGPT fare against a program specifically designed to play chess? Turns out, not well at all.
In 1979, Video Chess was released for the Atari 2600. The primitive console may be a relic by today’s standards, but it was a staple of many households in the late ’70s and early ’80s. Video Chess is so old that it can take minutes of real time for the computer to determine its next move on higher difficulty levels. If you thought loading times were bad on some Switch games, you didn’t play Video Chess back in the day. So what does ChatGPT end up doing when it goes against this primitive program? Lose terribly.
Robert Caruso is an engineer for Citrix software and he recently pitted ChatGPT against the Atari game. Apparently, Caruso was chatting with the AI program about the history of AI in chess. One thing led to another and next thing you know, the AI wants to play a round, wondering how quickly it could beat the game’s AI script. Instead, what proceeded to happen was ChatGPT complained about how the Atari Chess pieces looked. It also clearly did not understand what moves are legal and illegal.
To be fair, ChatGPT wasn’t designed with chess playing in mind. It’s not directed to learn the game, so pitting it up against something that was designed exclusively for that purpose was not going to be a fair fight. What would be even more interesting is to see what would happen if two ChatGPT’s were pitted against each other in a match. Or maybe the less we do to feed its self-learning, the better.
