Threat Assessment: High
An AI-generated image of actress Mia Goth, posted by a verified Instagram account with over a million followers, fooled celebrities, gossip blogs, and Google’s AI Overviews within days. The Verge AI reports that the fake showed Goth in a T-shirt reading “MY EX IS MY BIGGEST FAN,” timed just days after she filed for divorce from Shia LaBeouf. It racked up nearly 500,000 likes before Instagram pulled the account for impersonation.
This isn’t a one-off. It’s a pattern, and the systems meant to stop it are lagging behind.
How the Fake Spread
The chain of failure went like this:
- The account looked real. @miagoth_ had a blue checkmark. Meta sells those, even to accounts that aren’t official or run by public figures. The page called itself a fan account, but few people noticed.
- The timing sold it. The divorce news made the image feel plausible. Adding Beyoncé’s “Single Ladies” helped.
- Gossip outlets picked it up. Celebrity blogs treated the post as real news.
- Google amplified it. Search results showed an AI Overview saying Goth was taking “a clear swipe at her ex,” citing those same blogs. A featured section highlighted the “sassy” shirt.
What stands out here is the loop: an AI fake got laundered through human media and then repackaged as fact by a second AI system. Each step added credibility.
Goth’s representative, Christine Tripicchio, said they’d already complained to Instagram about this account posting “counterfeit, AI images.” She told The Verge:
“The fact that this fan account meets Instagram’s criteria to qualify for a ‘blue check’ is disconcerting on a fundamental level.”
The Wider Problem
The Verge’s reporting goes well beyond one post. Instagram is full of celebrity deepfake content:
- Sexualized content. One account posted only AI images of Scarlett Johansson and linked to Patreon and Telegram channels selling subscriptions. Another posts AI images of Millie Bobby Brown, Zendaya, and Ariana Grande, partly undressed or in sexual poses.
- Viral meme formats. An AI Taylor Swift rapping a Drake song, with explicit lyrics, has 1.8 million views. Identical clips feature Shrek (3.1 million), Lionel Messi (2.1 million), and Steve Jobs.
- Niche fan fiction. Some accounts exist only to show “Stranger Things actors getting married to each other.”
After The Verge flagged eight accounts, Meta removed some of them. Spokesperson Alex Dziedzan said Meta doesn’t allow accounts “to use someone’s likeness to pose as that individual, regardless of whether the content is authentic or AI-generated.” The Goth and Johansson accounts appear to be gone. The rest of the flood isn’t.
Context: How We Got Here
Celebrities have pushed back on AI clones since consumer tools first appeared:
- An AI Drake sound-alike track got pulled from YouTube after copyright strikes.
- Scarlett Johansson hired lawyers after OpenAI shipped a ChatGPT voice that sounded like her.
- Taylor Swift endorsed Kamala Harris in 2024, partly to correct fake images Donald Trump posted implying she backed him.
Those fights felt like early skirmishes over new norms. Now the volume has outrun the fights.
The Legal Gap
The US has no national right of publicity law. So likeness disputes mostly play out under each platform’s own rules.
There’s one exception. The Take It Down Act, signed in 2025, criminalizes sharing nonconsensual intimate imagery, AI-generated or not, and forces platforms to remove it within 48 hours. But that only covers intimate content. A fake T-shirt photo or a lip-sync video falls outside it.
YouTube has gone furthest on its own. Its likeness detection program scans for visual matches of enrolled people, and recently sound-alikes too, then lets them request takedowns. Not every request gets approved.
Tactical Takeaways
- Verification badges don’t mean authenticity anymore. Paid checkmarks tell you someone paid, not who they are.
- AI search summaries can repeat fakes. If blogs get fooled, Google’s overviews will cite them. Check the primary source.
- Brands and creators should watch their likeness. If it can happen to a Hollywood actress with a legal team, it can happen to a founder, an influencer, or your company’s spokesperson.
- Expect more pressure on platforms. Patchwork enforcement and missing federal law leave gaps that lawmakers and rights holders will keep pushing on.
Outlook
The tools that make these fakes keep getting cheaper and more convincing. Removal still mostly depends on someone complaining, and on a platform deciding to act. Until there’s a real legal standard for likeness, expect more cases where the fake spreads fast and the takedown comes late.
The full investigation is available at The Verge AI.