The Maia Chess engine uses open-source artificial intelligence software to accurately predict how humans of different skill levels play chess and points out mistakes a player should work on to improve their game.
The researchers who developed Maia hope that their work is a step towards creating better human-AI interactions in an increasingly digital and AI-heavy world.
Researchers converted how humans play and the mistakes they make into computer code to design Maia. The program played 12 million human games of varying levels.
Instead of trying to win each time, Maia tries to match human plays, accurately mimicking and predicting human moves and mistakes, helping players improve their game.
The work shows just how important the human element is when designing AI systems - they are easier to work with and learn from when built with human interaction, collaboration and improvement in mind.
Original article:
To err is not just human: U of T researchers develop AI that plays chess like a person
Original study:
Aligning Superhuman AI with Human Behavior: Chess as a Model System