Fake social media accounts have existed for as long as the platforms themselves. So when news broke recently that Emily Hart, a popular “hot girl” MAGA online personality, was actually a 22-year-old male medical student based in India, many might have written the story off as unremarkable. Just another catfishing scheme, another anonymous fake sock puppet, another internet scammer—there are thousands of them lurking online, after all.
But this case was far from ordinary. Unlike most small-time fakes, Emily had a full library of original photos and videos, amassed thousands of followers across multiple social platforms, and some of her posts racked up millions of views. She was a fully-fledged influencer, not a random anonymous troll account.
The man behind Emily admitted to Wired that while the account was active, he earned thousands of dollars each month sharing softcore videos on an OnlyFans alternative and selling branded merchandise. Surprisingly, he is not a professional tech developer or an organized scammer—just a cash-strapped student with a sharp understanding of U.S. political culture and a regular Google Gemini account.
Yet Emily Hart’s unusual case exposes a critical new reality: AI has made it shockingly simple for almost anyone to create hyper-convincing fake content and manipulate social media engagement systems. It also leaves us with urgent, unanswered questions: Are platforms or regulators doing enough to protect ordinary users? How can everyday people tell authentic content apart from AI-generated fakes anymore? And who bears responsibility for warning social media users when the media they are viewing was created with artificial intelligence?
The larger implication of this story is not about one single fake AI influencer. It is that Emily is only the tip of the iceberg. AI has lowered the barrier to building convincing fake online personas like Emily so drastically that large-scale deception is now accessible to almost anyone. Wired’s reporting already names other similar fake pro-Trump influencers, such as Jessica Foster, but you do not have to dig deep to find undisclosed AI-generated content on major platforms like Instagram. The Emily Hart case confirms that this blueprint for profiting from fake personas is cheap, fast, extremely lucrative, and simple for other bad actors to replicate.
All major social networks already have formal policies governing AI-generated content. While fine print varies across platforms, the core rule is generally consistent: synthetic content that could be mistaken for real must be clearly labeled as AI-generated, especially when it covers sensitive topics like politics, health, finance, or breaking news. Accounts that fail to disclose AI content can face penalties ranging from demonetization and frozen accounts to full bans.
But these penalties exist almost entirely on paper. In practice, enforcing these rules is incredibly difficult, in large part because detecting AI content gets harder every year. Today’s state-of-the-art AI image and video generators are light-years ahead of the early models that produced the viral, obviously fake “Will Smith eating spaghetti” clip. Common old tells of AI generation—like extra fingers or distorted background characters—are largely a thing of the past now. Without mandatory embedded watermarks to mark AI content, even automated detection tools struggle to tell the difference between an AI-generated image and a real photograph through visual inspection alone.
