Google has published a new patent for a system that would allow AI to handle game testing in the future.
Published on March 23, this new patent sees a system developed that would see AI actors perform gameplay actions in real-time. This new system called “EFFICIENT GAMEPLAY TRAINING FOR ARTIFICIAL INTELLIGENCE” would create a gameplay model for a character within a game, and would then train that character to perform the actions created.
Why Does Google Want AI Game Testing?
“Modern games are not just more complex than their predecessors, they reflect fundamental changes in the way games are designed and played,” Google says.
“Despite these dramatic shifts in the way games are played, the way that they are tested has remained largely unchanged,” the company continues. “As a result, game testing is a predominantly manual process, highly dependent on humans who repeatedly play the game and look for defects.”
Google says that the “complexity” of modern games still using traditional processes is leading to issues within studios. It believes that not moving forward is what is leading to games constantly being delayed as well as games launching in poor states.
With its system, predefined criteria would be set to make sure games are tested properly. It’s also mentioned that the model can be changed and updated based on “additional observational data” based on how a human user plays the game. That would suggest that people would still play a role in game testing in some way, especially when it comes to performing actions that developers and, thus, the AI wouldn’t have thought of initially.
What do you think about Google’s thoughts on AI handling game testing in the future?
Want more Insider Gaming? Check out the latest episode of The Indie Initiative featuring Temple of Starlight from Enlit Games.