October 7th, A TikTok account named @fujitiva48 asked a provocative question along with its latest video. “What do you think about this new toy for little kids?” they asked more than 2,000 viewers who stumbled across what appeared to be a parody of the TV commercial. The response was clear. “Dude, this is not funny,” one person wrote. “There should be an investigation to find out who made this.”
It's easy to see why this video sparked such a strong reaction. The fake commercial begins with a photorealistic young girl holding a pink, sparkly toy with a bumblebee emblazoned on the handle. We can tell that it's a pen because the girl and two others are scribbling on paper while an adult male narrator narrates the story. But the object's floral pattern, ability to make buzzing noises, and name Vibrorose make it clear that it looks and sounds very similar to a sex toy. The “Add” button, a TikTok feature that prompts people to share a video on their feed along with the words “I use rose toys,” removes even the slightest doubt. (WIRED reached out to the @fujitiva48 account for comment, but received no response.)
This disturbing clip was created using OpenAI's latest video generator, Sora 2. Sora 2 was first released on an invite-only basis in the US on September 30th. Within just a week, videos like the Vibro Rose clip migrated from Sora and landed on TikTok's For You page. Other fake ads were even more blatant, and WIRED found several accounts posting similar Sora 2-generated videos featuring water toys shaped like roses and mushrooms, and cake decorators squirting “sticky milk,” “white foam,” or “goo” onto realistic-looking images of children.
In many countries, the above would be the basis for an investigation if these were real children rather than digital fusions. However, laws regarding AI-generated fetish content involving minors remain vague. New data from the UK's Internet Watch Foundation for 2025 shows that the number of reports of AI-generated child sexual abuse material (CSAM) doubled from 199 in the year from January to October 2024 to 426 in the same period in 2025. 56 percent of this content falls into category A. 56 per cent of this content falls within the UK's most serious categories, including invasive sexual activity, sex with animals or sadism, and 94 per cent falls within category A. One of the illegal AI images tracked by IWF was of a young girl. (Sora does not seem to generate content for Category A.)
“While we often see portraits of real children being commercialized to create nude or sexualized images, we overwhelmingly see AI being used to create images of young girls. This is another way girls are targeted online,” IWF chief executive Kelly Smith told WIRED.
The influx of AI-generated harmful substances has prompted the UK to introduce new amendments to the Crime and Policing Bill, which would allow “certified testers” to check whether artificial intelligence tools are unable to generate CSAM. As reported by the BBC, the amendments will ensure that models have safeguards in place when it comes to certain images, particularly extreme pornography and non-consensual intimate images. As AI-generated programs continue to evolve, 45 states in the United States have enacted laws criminalizing AI-generated CSAM, most within the past two years.
