Lifestyle

School photos in the UK are being weaponised by criminals with AI – but the government won’t say how exposed children are

Ryan Brothwell 4 min read
School photos in the UK are being weaponised by criminals with AI – but the government won’t say how exposed children are

Key Points

  • School photos are being scraped to generate AI child abuse imagery, the IWF and NCA warned in May.
  • At least one UK secondary school was blackmailed with AI fakes built from images on its website.
  • Lib Dem MP Liz Jarvis asked the DfE what it had assessed; the reply cited law and guidance, not an assessment.
  • The government has published no specific measure of how exposed children are to school-photo scraping.
  • New AI-tool offences sit in the Crime and Policing Act 2026, not the Online Safety Act the minister cited.

Criminals are scraping publicly posted school photographs and feeding them into artificial intelligence tools that generate child sexual abuse imagery, then using the results to extort.

In May, the Internet Watch Foundation and the National Crime Agency warned schools to remove or obscure pupil images after at least one unnamed secondary school was targeted in a blackmail attempt, with offenders lifting photos from its website and social accounts, producing explicit fakes and demanding payment.

The IWF classed 150 of those images as illegal under UK law, and said it knew of other attempts of the same kind.

So when an MP asked the government how exposed children actually are, it was a fair question. The answer, published on 2 June, did not give a number.

The question and the non-answer

Liz Jarvis, the Liberal Democrat MP for Eastleigh, asked the Department for Education what assessment it had made of the risk posed by criminals using AI to turn publicly available school photographs into abuse imagery. It is a precise question about a documented threat.

The reply, from education minister Josh MacAlister, contains no assessment.

What it contains is a tour of the statute book. The law is clear, the minister wrote, that producing, storing, sharing or searching for such material is an offence whether or not it is AI-generated.

The Online Safety Act requires platforms to remove illegal content quickly.

The “Keeping children safe in education” statutory guidance sets expectations, and will be strengthened in its 2026 edition with cyber security content and signposting to the National Crime Agency’s CEOP guidance on sextortion.

There is also, he noted, separate guidance on taking and using photos in schools.

What Whitehall will say

The Crime and Policing Act 2026 received Royal Assent last month, and it made Britain the first country to criminalise the possession, creation or distribution of AI tools built to generate child abuse material, alongside an offence covering AI models optimised to produce hyper-realistic imagery carrying the likeness of real children.

HotMinute has tracked the Online Safety Act’s expansion as ministers fast-tracked online harms reforms through the spring.

MacAlister credits the Online Safety Act with introducing the new offences against AI generation tools. Those offences in fact sit in the Crime and Policing Act, a separate regime that completed its passage only weeks ago.

The slip is minor. It is also characteristic of a reply that describes the architecture of the law rather than the specific danger it was asked about.

Jarvis did not ask what the law says. She asked what the Department had assessed. On that, the answer is silent.

There is no published figure, no scoping of how many schools may be exposed, no estimate of how widely pupil images circulate, and no sign that the scraping risk has been examined as a distinct problem rather than folded into general safeguarding guidance.

This is particularly concerning as the IWF’s latest research found AI abuse material accelerating, with most of the AI abuse videos it identified last year falling into the most serious legal category. Against that, the guidance the minister points to is thinner than it sounds.

The data protection material on photographs is built around consent and lawful processing, the questions schools have wrestled with for years.

It was not written for a world in which a lawfully posted sports-day photo becomes a stranger’s source file. The sharper advice, to favour distant or obscured shots and to ask whether a photo is needed at all, comes from the advisory Early Warning Working Group rather than the Department.

None of this means ministers are concealing anything. Written answers are short by design, and the legal response is real.

But a parent who reads this exchange, learns what is illegal and where to find guidance, will find this all cold comfort.

Now read: How a Chatbot became Britain’s new favourite comparison site