A playable AI game is not a finished game
WIRED’s Boone Ashworth tried making games with Google’s AI Playground. The games ran, but one lost its intended theme and another had confusing controls. A later fix could bring a new problem. A preview shows what the tool made once; it does not show that a game is ready for players. This was one writer’s test, not a study of all AI-made games.
What the source reports
WIRED’s Boone Ashworth tested Google’s AI Playground by making games. The games were playable, but a working first version was not the same as a good game. In one build, the intended theme did not carry through. In another, the controls were hard to understand. Fixing one flaw could introduce a new one. The account gives readers a close look at a product test and its revisions. It does not measure how often the tool succeeds across users or tasks. The test leaves a clear question: does the game still work as intended after people try it and the maker changes it?
Original source
- Title
- I Made Terrible Games With Google’s AI Playground
- Author
- Boone Ashworth
- Date
- Sunday, October 11, 2026