You probably know isn’t the kind of video game that encourages adherence to posted speed limits. Actually I’m pretty sure it’s impossible to play without committing multiple virtual felonies in the time it takes to load the title screen. So why is it being used as a training platform for autonomous cars?

Of course there is more to the game than heists and wanton violence. There’s a massive and realistic-looking 3D working world, and that’s what research groups like Intel Labs and Germany’s Darmstadt University are interested in.

As MIT’s reports, GTA V is being used to train artificial intelligence mainly to recognize objects—cars, pedestrians, bicycles, buildings, roads and other things a car might encounter in the real world.

“The researchers created a software layer that sits between the game and a computer’s hardware, automatically classifying different objects in the road scenes shown in the game. This provides the labels that can then be fed to a machine-learning algorithm, allowing it to recognize cars, pedestrians, and other objects shown, either in the game or on a real street.”

The short story is that data that would take an eternity for a camera-car driving around a real city to collect and people to subsequently label is instantly available in the game.

It’s really just giving software developers a lot more information to work with when trying to train an autonomous car how to function in the real world. Not teaching cars to raise hell and rack up . Which would be much more hilarious, but, problematic.

Commenter cuts_off_prius found some videos of similar software in action. I originally thought this was the same stuff discussed in the MIT study, but another reader Kakurady tells us out that this video showcases “” which actually uses screenshots to “learn” from.

This “DeepDrive” is apparently unrelated to , which is also studying artificial “perception” for vehicles.

As for GTA being used in the studies discussed previously–

“With artificial environments we can effortlessly gather precisely annotated data at a larger scale with a considerable amount of variation in lighting and climate settings,” PhD student at the University of British Columbia Alireza Shafaei, who has studied training computers with video games, told Tech Review. “We showed that this synthetic data is almost as good, or sometimes even better, than using real data for training.”

Still, I feel like an ill-intentioned prankster could get in there and be a bad influence on the cars of the future. But I guess you could say that about .