Illustration by Annelise Capossela/For The Washington Post; iStock
Last year, Dr. Joel Bervell, who has built a social media following as a medical “mythbuster,” found an AI version of himself selling a product he had never used on YouTube. “As a physician, that was especially alarming because they weren’t just borrowing my image,” he said. “They were leveraging the trust that comes with my medical license to make it appear that I was recommending something to my audience.”
Stories like Bervell’s were the pitch when more than a dozen voice actors and creators, including myself, joined an event organized by YouTube to announce the new Senate Creator Caucus and bring awareness to the spread of deepfakes and AI-generated content that mimics real people’s name, image and likeness. Actors, creators, and major social media companies, like YouTube, asked lawmakers to pass the No Fakes Act.
To be upfront, I’ve been a creator for years, and I personally believe Congress needs to rein in AI deepfakes on social media, even though this particular legislation deserves some scrutiny. Here's the gist: The bipartisan bill would give every American a federal right to control what it calls a “digital replica” of themselves: an unauthorized AI-generated copy of their voice or visual likeness. That could cover voice cloning and other impersonation that may not fit neatly under existing fraud or scam laws.
Led by Sens. Chris Coons, D-Del., and Marsha Blackburn, R-Tenn., the bill cleared the Senate Judiciary Committee unanimously in June, but the window for passage is narrowing.
If enacted, anyone who sees an unauthorized use of themselves on social media would report it to the social media platform where they found it to ask for it to be taken down and removed. Platforms that comply with the notice-and-takedown process would receive liability protection, while those that knowingly fail to remove unauthorized replicas could face penalties of up to $750,000. Individuals can face $5,000 in penalties per work, and other companies up to $25,000 per work, or potentially actual damages and profits if those are greater.
Certain uses in news, documentaries, commentary, satire and parody would be exempt. Supporters say those exemptions and the bill’s counter-notice process are designed to protect First Amendment rights, but opponents say it doesn’t go far enough.
Rep. Madeleine Dean, D-Pa., a sponsor of the House bill, said it would “empower victims of deep fakes” and that “we must support technological innovation while preserving the privacy, safety, and dignity of all Americans.”
The No Fakes bill aims to go a step beyond the Take It Down Act, passed May 19, 2025, after advocacy from sexual violence survivors and public figures (including Paris Hilton) claimed that nonconsensual intimate images and sexually explicit deepfakes were harmful and needed regulation. The No Fakes bill proposes to cover unauthorized replicas of a person’s voice and likeness more broadly.
For creators, a convincing fake post can cost them sponsorships, audience trust and credibility. Therapist and creator Kati Morton said she encounters many fake versions of herself each day. Viewers may be manipulated by bad actors spreading misinformation or tricked into buying a product from their favorite creator, only to discover that the endorsement was a deepfake scam.
Reuters reported that Meta internally projected about $16 billion, or 10% percent of its 2024 revenue, would come from ads for scams and banned goods, though Meta disputed the estimate. No Fakes would not impose requirements for social media platforms to find every unauthorized replica before it is reported. But after receiving a valid notice, user-upload platforms such as Meta would have to remove the identified material and future uploads matching its digital fingerprint to qualify for the bill’s safe harbor.
“You shouldn’t have to be famous to deserve control over your own identity,” said Stephanie Patrick, who sold her multi-channel YouTube studio Theorist Media with her creator husband, MatPat, in 2022. “As AI makes it easier to replicate anyone’s face or voice, the potential for impersonation and exploitation extends to all of us, from famous personalities to underage kids in school.”
Still, many free speech and digital rights groups oppose the bill or want significant changes, saying it could harm free speech or lets social media platforms off the hook for liability of harmful content. In a letter to lawmakers led by the Center for Democracy & Technology, the ACLU, Electronic Frontier Foundation, Fight for the Future, Public Knowledge and Woodhull Freedom Foundation warned that the bill’s notice and takedown system could encourage platforms to remove lawful speech rather than risk liability.
"For example, a satirical video that uses an AI-generated version of a politician's image to criticize a law could be taken down after a complaint, even though the video is clearly a parody," said Tess Vartanian of CDT.
John Coleman, legislative counsel for the group Foundation for Individual Rights and Expression, flagged the bill's subpoena provisions, which would let a court order platforms to identify anonymous posters. While some anonymous accounts may be bad actors, they "could be used to unmask anonymous critics or whistleblowers before a court even determines their speech is unlawful," he said.
The reality check for the legislation is timing. The bill has no scheduled Senate floor vote, while the House version has not cleared committee. Supporters hope for a vote during the post-election lame-duck session, potentially as part of a year-end package.
Meanwhile, platforms that support the legislation are building tools. TikTok is testing what it calls “Likeness Detection,” which enables creators to review and report flagged AI content. YouTube says its technology similarly helps creators identify altered or AI-generated videos and request removal.
Startups are entering the market, too. Vermillio founder Dan Neely says his company helps creators detect unauthorized AI outputs and manage consent, attribution and compensation. He estimates that unauthorized content can consume 5% to 25% of a creator’s audience.
Bervell supports the bill’s goal, but emphasized that lawmakers must protect legitimate journalism and creative expression. "We should be able to protect people from AI-enabled impersonation without undermining the ways people legitimately speak about, depict, and critique one another," he said.
No solution will be perfect. Congress, platforms and all of us have to keep refining the safeguards, but at some point we must decide whether the risks of moving forward outweigh the growing cost of doing nothing.
Disclosure: Lazar’s trip to Washington earlier this month was sponsored by YouTube, where she joined several creators to attend an event announcing the Senate Creators Caucus at the U.S. Capitol and a day of meetings with lawmakers, where participants advocated for the No Fakes Act. She also has previous paid partnerships with YouTube.
This story is part of Verified, a newsletter that is published by Washington Post Creator, a team outside The Washington Post’s newsroom that is focused on the creator economy and content partnerships with independent creators. Learn more about Washington Post Creator.
