Video Game

Driverless Cars Are Learning How Not To Kill Us From Grand Theft Auto

Video games can be purposeful for real life.

Autonomous car developers have a problem, and one of the only ways to solve it is with Grand Theft Auto V. The problem, according to Bloomberg, is that there’s not enough hours in a day for driverless car prototypes to teach themselves how to drive in the real-world. Video games, such as GTA, are proving to be a helpful workaround. Software developers like this mega popular video game because it provides a simulation platform that can "generate data that’s very close to what artificial-intelligence agents can glean on the world."

Given that companies like Ford and Google’s Waymo are in a race to get driverless cars on the road in the next three years or so, having such a virtual reality platform is essential – and it’s readily available at very little cost. Developers like GTA because it has the ability to generate real-life situations, often times dangerous, that can be hard, or even impossible, to replicate in a real-world scenario. The self-driving software needs to experience certain situations, like a mattress falling off a vehicle on a highway, in order to learn how to respond quickly and, above all, safely.

“Just relying on data from the roads is not practical,” states David Bacchet, a driverless EV software team leader from San Jose, California. “With simulation, you can run the same scenario over and over again for infinite times, then test it again.” Sounds a little bizarre at first, we know, but the logic is clearly there. In fact, GTA isn’t the only video game developers have used for go-to AI software. Super Mario Bros. and Angry Birds have also proven capable of simulating real-world problems in virtual reality environments.

Read Next

SEE MORE ARTICLES