AI-generated sexual images of teachers, made by their own students, are surfacing in classrooms across the United States and the United Kingdom, and schools have no consistent playbook for responding. Four educators told their accounts of being deepfaked by students, including Luis DeSantiago, a substitute teacher in Los Banos, California, who found a manipulated lingerie image of himself circulating on TikTok on March 12. The federal Take It Down Act makes such nonconsensual synthetic imagery a crime, but enforcement against minor students is uneven, and the platforms hosting the content are inconsistent responders.
DeSantiago learned about the image from a co-worker at his second job. The picture appeared to be built on top of a real mirror selfie he had taken years earlier in college, altered to place him in lingerie and a suggestive pose. He traced the initial TikTok account to a student at the local middle school where he regularly substituted.
He reported the image to the assistant principal the same evening. That night he also discovered a second account impersonating him, using his name with a slight misspelling and the same profile photo.
“When I finally saw the image, that's when it felt like a punch to the gut.”— Luis DeSantiago, substitute teacher, Los Banos, California
Key facts
- 01Luis DeSantiago, a substitute teacher in Los Banos, California, discovered an AI-generated lingerie image of himself circulating on TikTok on March 12.
- 02The federal Take It Down Act makes it a crime to publish nonconsensual intimate depictions online, including AI-generated ones.
- 03A UK college counselor named Alice, who mentors roughly 150 students per year, was targeted by a manipulated video showing her undressing.
- 04One student who admitted creating a sexual deepfake of DeSantiago received only in-school suspension; the image continued to spread via Snapchat and AirDrop.
- 05Attorney Donna Ballman says student-made sexual images of teachers constitute a clear case of workplace sexual harassment under federal law.
One student later admitted to creating the image and received in-school suspension. The image continued to spread anyway — it appeared on Snapchat, and when DeSantiago took an assignment at the local high school, someone AirDropped it to him mid-class before quickly revoking the send. He told administrators he would not return to the middle school until the matter was fully addressed, and told WIRED he probably will not return to the district this year.
The legal architecture around this behavior exists but is thin at the school level. The Take It Down Act criminalizes publishing nonconsensual intimate depictions, both authentic and computer-generated. Snap spokesperson Monique Bellamy pointed to company policies prohibiting AI-generated or manipulated content used to humiliate or sexually harass others, and said Snap has established reporting processes. TikTok did not comment before publication.
Donna Ballman, a retired employment advocacy lawyer, said a student creating and sharing sexual images of a teacher is a clear case of workplace sexual harassment, and federal antidiscrimination law obliges employers to maintain a harassment-free workplace. But she said schools have no blueprint for how to respond, and administrators are not required to tell a victimized teacher how a student was disciplined.
“They just have to make [and] create a safe environment for the victim.”— Donna Ballman, retired employment advocacy lawyer
A middle school teacher in her mid-twenties, who requested anonymity, described a pattern rather than an isolated event. Last year, students at her school manipulated public photos of a teacher in her sixties to depict sex acts and shared them in a Snapchat group. A few months ago, a student created a TikTok account impersonating her, with AI-generated video showing her mouth moving unnaturally while describing sexual acts.
She said students reported the account to her throughout the day, and many believed it was real. She asked them to report it to TikTok, and the account was taken down. At the next faculty meeting a colleague reported a similar AI-generated video of himself getting arrested, which he believed came from the same student.
The harassment then escalated in person. The same student began cursing her out and calling her names, and was texting and calling the assistant principal in the middle of the night while posting AI-generated photos of the administrator to TikTok. The student eventually left the district. She said she worries about a repeat, noting that the behavioral profile is not unique to one student.
Alice, a college counselor at a UK trades school for students aged 16 to 19, meets individually with roughly 150 students a year on topics ranging from drugs and safe sex to academic issues. This winter, administrators told her two videos were circulating — one real conversation with a student, and a manipulated version showing her removing her clothes and dancing in what appeared to be underwear or a bikini. The student she suspected denied involvement and dropped out.
The incidents point to a widening gap between what generative tools now make trivial and what schools, platforms, and law enforcement are equipped to address. Ballman's point cuts to the operational reality: the law requires a safe environment for the victim, but does not tell administrators how to build one when the perpetrator is 13 years old and the distribution surface is every phone in the building. Consumer-grade image and video models have collapsed the skill barrier to producing convincing fake nudes, and the enforcement layer — school discipline, platform takedowns, criminal referrals — was designed for a slower, higher-friction world.
For the AI industry, this is the class of harm that policy debates about frontier risk keep skipping over. It doesn't require a superintelligent model or a novel jailbreak. It requires a mid-tier image editor, a public photo, and a group chat. The Take It Down Act gives victims a federal cause of action, but the practical remedy for a teacher whose face is on a lingerie photo passed around a school AirDrop is still measured in weeks of administrative back-and-forth. Until platforms treat synthetic sexual content targeting identifiable people as a hard-block category rather than a reactive-takedown one, the caseload will keep growing, and the people absorbing it will be the ones with the least ability to opt out — teachers, coaches, and coworkers whose faces are already public.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
