AI nonprofit TrueMedia gets real with release of free deepfake detection tool ahead of elections

TrueMedia, the Seattle-based nonpartisan nonprofit that is using AI to detect deepfakes and combat disinformation, made its technology available to the public on Tuesday.

The no-cost, web-based tool, which was first released earlier this year to journalists, fact-checkers and others, is now available to everyone ahead of the U.S. elections. Users can share a social media post containing an image, video or audio file and TrueMedia’s AI, with the help of existing deepfake detection tools, will analyze the content in real time for evidence of manipulation.

“For the first time, the public has easy access to the same best-in-class deepfake detection technology historically reserved for government agencies,” said Oren Etzioni, a longtime computer scientist, AI specialist and the founder of TrueMedia. “In an election cycle where disinformation is rampant, it is critical that everyone has the tools they need to verify the authenticity of what they see and hear online.”

Pop star Taylor Swift referenced a fake post in which AI was used to make it look like she was endorsing Donald Trump for president. As she endorsed Vice President Kamala Harris in an Instagram post last week, she said the prior experience conjured up her “fears around AI and the spreading of misinformation.”

In a news release, TrueMedia shared other examples of deepfakes its tech helped identify during major global events, including:

• A fake image of smiling Secret Service agents during the July assassination attempt on former President Trump.
• A fake Bollywood party endorsement during the 2024 Indian general elections.
• Russian state media broadcast of a fake audio clip of a Ukrainian Defense Minister taking credit for the Crocus Center bombing in March 2024.

TrueMedia said it also helped identify 41 AI content farm accounts that published 9,784 videos, from March 2023 to June 2024, cumulatively garnering over 380 million views.

In June, TrueMedia released a quiz to test people’s “deepfake IQ” and how well they can spot fake images, videos and audio clips.

Etzioni, a University of Washington professor and former CEO of the Allen Institute for AI, has called widespread deepfakes used to manipulate voters “a kind of disinformation terrorism” and said a tactic once deployed by state actors is now, unfortunately, something anyone can do.

TrueMedia announced a collaboration earlier this year with Microsoft’s AI for Good Lab, to enhance AI deepfake detection capabilities.

“Deepfakes are getting harder to spot,” Microsoft President Brad Smith said, calling TrueMedia’s tools “a great example of using good AI to combat bad AI.”

Etzioni will speak about AI, misinformation and more in an event at Startup Hall at the University of Washington in Seattle on Sept. 26.