Adobe unveils authentication tool in battle against deepfakes
Adobe Inc. debuted a software tool to help media creators prove their images are real, the latest move by the maker of Photoshop to combat the spread of deepfake technology.
An attribution tool for Photoshop and Behance, a social-media service owned by Adobe, will be available for testing in the coming weeks and generally released in 2021, the San Jose, California-based company said Tuesday in a statement. The software feature will let creators tag pictures with their names as well as the history and location of edits, to provide more transparency to a public growing increasingly skeptical of digital images.
Adobe is undertaking this project as part of the Content Authenticity Initiative, a coalition of technology and media companies including Microsoft Corp., Twitter Inc., the British Broadcasting Corp. and the Canadian Broadcasting Corp. These partners hope to build a new, open-source digital standard to help consumers online distinguish between genuine images and misinformation. Microsoft last month released a tool to help analyze videos and photos and provide a score indicating the chance that those images have been manipulated.
The coalition's effort is meant to build trust for images that feature transparent data and be more distrustful of pictures of unknown provenance. Adobe doesn't generally weigh in on political matters, but Dana Rao, the company's executive vice president, general counsel and corporate secretary, said he feared Americans would "disengage" from the political process due to online disinformation.
"We're worried that if people lose faith in facts that we aren't going to be able to have any conversations about moving the country forward," Rao said in an interview. "If none of us can agree on what is actually happening, we can't have a policy discussion."
Rao said he wishes he could go back in time five years and undertake the initiative then, but believes the technology may help in the elections in two and four years.
Adobe's photo-editing products have made image manipulation mainstream. Photoshop and Lightroom are ubiquitous applications used to airbrush magazine covers, create art, and in some instances, to misinform the public. Rao said that in 99% of cases, customers use Adobe's software in "productive, beautiful ways."
Rao said Adobe's tool doesn't try to detect deepfakes because artificial intelligence will make it increasingly difficult to keep up with and identify those fraudulent images. Instead, the company is trying to help consumers avoid manipulation by authenticating real digital images. While Adobe's Photoshop engineers sometimes help law enforcement examine images of child abduction or exploitation for photo manipulation, the company doesn't have enough people to quickly evaluate and call out a wave of political disinformation about the Nov. 3 U.S. election.
"This is what we're referring to internally as our digital citizenship," Rao said of the effort against deepfakes. "We are committed to ensuring our technology and the use of our technology benefits society."
©2020 Bloomberg News
Distributed by Tribune Content Agency, LLC.