Adobe AI can spot facial manipulations in Photoshop

Adobe and UC Berkeley say their artificial intelligence tool can spot doctored images 99% of the time.

Corinne Reichert Senior Writer
Corinne Reichert (she/her) grew up in Sydney, Australia and moved to California in 2019. She holds degrees in law and communications, and currently writes news, analysis and features for CNET across the topics of electric vehicles, broadband networks, mobile devices, big tech, artificial intelligence, home technology and entertainment. In her spare time, she watches soccer games and F1 races, and goes to Disneyland as often as possible.
Expertise News, mobile, broadband, 5G, home tech, streaming services, entertainment, AI, policy, business, politics Credentials
  • I've been covering technology and mobile for 12 years, first as a telecommunications reporter and assistant editor at ZDNet in Australia, then as CNET's West Coast head of breaking news, and now in the Thought Leadership team.
Corinne Reichert
2 min read

Photoshop is being used to change people's facial expressions.

Alexandra Robinson/AFP/Getty Images

In a world filled with manipulated photos, deepfakes and even totally fake human faces, Adobe says it's working on an artificial intelligence tool to spot fake images. Citing "the ethical implications" of Photoshop , Adobe partnered with researchers from the University of California at Berkeley to work on the issue.

Photoshop's Face Aware Liquify feature is being used to change people's facial expressions, the company said.

"Fake content is a serious and increasingly pressing issue," Adobe said in a blog post Friday, adding that it will use AI to increase trust in digital media. 

Adobe and Berkeley researchers have now developed a way to detect and remove edits to images.

Their tool was able to spot altered faces 99% of the time, in comparison to the human eye, which found the alterations 53% of the time, Adobe said. It was also able to revert images to their original state.

The tool is still in its early stages, however.

"The idea of a magic universal 'undo' button to revert image edits is still far from reality," Adobe researcher Richard Zhang said. "But we live in a world where it's becoming harder to trust the digital information we consume, and I look forward to further exploring this area of research."


Adobe's AI tool for spotting doctored images.


Manipulated images and videos have been spreading across the internet, like Friday's video that saw Jon Snow apologize for Game of Thrones season 8, and reports Thursday that a spy set up a fake LinkedIn account with an AI-generated photo to connect with political figures in Washington.

Deepfakes, video forgeries that make people appear to be doing or saying things they didn't, are the moving-picture equivalent of bogus images created with programs like Photoshop. Deepfake software has made manipulated videos accessible and increasingly harder to detect as fake. One technique enables users to make a deepfake using a single image, such as the Mona Lisa.

Congress is looking to investigate deepfakes following the appearance of doctored videos of US House Speaker Nancy Pelosi and amid fears that deepfakes could escalate the fake news campaign during the 2020 US presidential race.

Watch this: Senate takes on deep fakes with Sheryl Sandberg and Jack Dorsey

As with Adobe's tool for spotting doctored images, researchers are also working on a tool that detects deepfakes of leaders by creating pseudo-fingerprints of their unique manners of speaking. The new technique was outlined in an academic paper on Wednesday.