Skip to Content
Artificial intelligence

Adobe’s new AI tool can spot when a face has been Photoshopped

Altered images of the same face
Altered images of the same faceAdobe

It was nearly twice as good at identifying manipulated images as humans.

The research: Researchers from Adobe and UC Berkeley have created a tool that uses machine learning to identify when photos of people’s faces have been altered. The deep-learning tool was trained on thousands of images scraped from the internet. In a series of experiments, it was able to correctly identify edited faces 99% of the time, compared with a 53% success rate for humans.

The context: There’s growing concern over the spread of fake images and “deepfake” videos. However, machine learning could be a useful weapon in the detection (as well as the creation) of fakes.  

Some caveats: It’s understandable that Adobe wants to be seen acting on this issue, given that its own products are used to alter pictures. The downside is that this tool works only on images that were made using Adobe Photoshop’s Face Aware Liquify feature.

It's just a prototype, but the company says it plans to take this research further and provide tools to identify and discourage the misuse of its products across the board.

This story first appeared in our daily newsletter The Download. Sign up here to get your dose of the latest must-read news from the world of emerging tech.

Deep Dive

Artificial intelligence

A Roomba recorded a woman on the toilet. How did screenshots end up on Facebook?

Robot vacuum companies say your images are safe, but a sprawling global supply chain for data from our devices creates risk.

The viral AI avatar app Lensa undressed me—without my consent

My avatars were cartoonishly pornified, while my male colleagues got to be astronauts, explorers, and inventors.

Roomba testers feel misled after intimate images ended up on Facebook

An MIT Technology Review investigation recently revealed how images of a minor and a tester on the toilet ended up on social media. iRobot said it had consent to collect this kind of data from inside homes—but participants say otherwise.

How to spot AI-generated text

The internet is increasingly awash with text written by AI software. We need new tools to detect it.

Stay connected

Illustration by Rose Wong

Get the latest updates from
MIT Technology Review

Discover special offers, top stories, upcoming events, and more.

Thank you for submitting your email!

Explore more newsletters

It looks like something went wrong.

We’re having trouble saving your preferences. Try refreshing this page and updating them one more time. If you continue to get this message, reach out to us at customer-service@technologyreview.com with a list of newsletters you’d like to receive.