There may come a time when there won’t be a story about someone using artificial intelligence (AI) in a horrific manner to do something illegal, immoral, or just gross. Today is not that day. In fact, this one involves all three.
Federal prosecutors in the Eastern District of New York unsealed a criminal complaint on May 19, 2026, charging 20‑year‑old Arturo Hernandez of Bedias, Texas, in connection with AI‑generated sexual “deepfake” content. The complaint alleges that Hernandez used AI tools to create explicit images and videos from non‑sexual photos and then posted them online in violation of a relatively new federal law referred to as the Take It Down Act, which targets nonconsensual digital forgeries of intimate imagery. According to prosecutors, Hernandez’s content depicted roughly 50 women, including public figures and women he knew personally, none of whom gave permission for their images to be used.
The case is a grim reminder of how quickly technology can be weaponized. A decade or two ago, you couldn’t ruin someone’s life by feeding their graduation photos into an algorithm and hitting “render.” Now you can, and you can face serious federal charges for doing it.
Maybe Don’t Put It Up in the First Place
Passed by Congress with unilateral bipartisan support (409-2 in the House, unanimous in the Senate) in Spring, 2025, and signed into law a month later, the Take It Down Act created federal law for, among other aspects, the posting of nonconsensual or digitally created sexual images and videos. The new law adds a federal criminal remedy focused specifically on AI‑generated and nonconsensual intimate imagery, going beyond earlier civil and state‑level approaches to revenge porn such as the Violence Against Women Act (VAWA).
The statute also fits into a wider policy conversation about how far Congress can go in nudging platforms to respond to intimate‑image abuse while still respecting existing protections for online intermediaries. It encourages online services to act on reports of nonconsensual explicit content and deepfakes, while generally preserving the familiar immunity framework for user‑generated content. What it does not do is make every platform liable for every bad thing a user posts; instead, it creates a new criminal hook for the people who create and distribute the deepfakes in the first place.
Whether you consider it an exciting new tool that lets anyone become a filmmaker or an unfortunate side effect of opening the AI Pandora’s Box, modern tech makes it easy to turn a static image into something that looks like a scene from a movie. In Hernandez’s case (and another person indicted for similar offenses at the same time), the desire to create was focused on a very specific genre. (Porn. It’s usually porn.)
Less a Trail of Breadcrumbs and More a Bunch of Arrows Pointing at Texas
Despite Hernandez’s use of anonymous emails and account names, the Federal Bureau of Investigation (FBI) was able to track him down in Texas. In addition to using similar handles for his accounts, Hernandez identified himself by name in an Instagram account that included an email address associated with his uploads. The FBI was also able to match the IP address used to post the illegal materials to the town where Hernandez went to high school.
Despite his posting of images and videos of over 50 people, the arrest singles out nine specific victims (who are not named). That’s still very bad news for Hernandez. Under the Take It Down Act, each deepfake image or video carries a potential sentence of two years in a federal prison. The criminal complaint names 57 such items, with hundreds more mentioned but not part of the arrest warrant.
Hernandez’s postings received hundreds of thousands of views before being taken down.
