How deepfake images could be key to teaching autonomous cars how to drive

The AI-generated images are helping to train autonomous driving software 
Oxbotica's deepfake technology can apply an infinite number of scenarios to a particular scene
Oxbotica
Amelia Heathman25 June 2020

Deepfake images are one of the darkest parts of the internet — the AI generated images can be used to place celebrities in porn films or doctor what a politician said. But the tech does have its uses.

The autonomous driving pioneer Oxbotica has been using deepfake tech to train its software on different scenarios without having to put self-driving cars through them in real-world locations.

Thousands of photo-realistic images can be created in minutes, offering up infinite variations of the same situation, such as reproducing the same scene but with poor weather or adverse conditions. Oxbotica says the tech is so advanced it can change the lighting of an image, or replace one object in an image, such as a tree, with another, like a building. These scenarios are then used to teach the autonomous software which will later be used in vehicles.

Oxbotica’s co-founder and CTO Paul Newman said using deepfakes is an “incredible opportunity” for autonomous vehicles, given that it can increase the speed and efficiency of training a vehicle.

Oxbotica has been testing autonomous cars in places such as Hounslow

“What we’re really doing here is training out AI to produce a syllabus for other AIs to learn from. It’s the equivalent of giving someone a fishing rod rather than a fish. It offers remarkable scaling opportunities,” explained Newman.

In the autonomous vehicle space, companies can get hung up on the amount of miles the vehicles have travelled as a way to showcase their expertise. But miles travelled doesn’t necessarily mean the software has encountered all the different situations it would face when it is later carrying passengers. This is where the use of the deepfake tech can come in.

“The use of deepfakes enables us to test countless scenarios, which will not only enable us to scale our real-world testing exponentially; it’ll also be safer.”

The tech world is very cautious when it comes to deepfake tech. The social network Facebook banned manipulated videos on the platform earlier this year, calling them a “significant challenge for our industry and society as their use increases.” Yet the Oxbotica example shows that AI-generated images can be used for good — the difficulty is ensuring that this is all it's used for.

Create a FREE account to continue reading

eros

Registration is a free and easy way to support our journalism.

Join our community where you can: comment on stories; sign up to newsletters; enter competitions and access content on our app.

Your email address

Must be at least 6 characters, include an upper and lower case character and a number

You must be at least 18 years old to create an account

* Required fields

Already have an account? SIGN IN

By clicking Create Account you confirm that your data has been entered correctly and you have read and agree to our Terms of use , Cookie policy and Privacy policy .

This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.

Thank you for registering

Please refresh the page or navigate to another page on the site to be automatically logged in