Why AI is every teacher's responsibility

Every September, every member of staff in every school sits down for safeguarding training, quite rightly, because Keeping Children Safe in Education (KCSiE) is clear that safeguarding is everyone's business. Nobody thinks it can be left to the designated safeguarding lead alone.

This year, something new appeared in that training. For the first time, KCSiE refers explicitly to artificial intelligence. That got me thinking. If AI now has a place in guidance every member of staff must know, shouldn't AI training itself be mandatory for every member of staff?

I think it should. The most immediate reason is the one KCSiE itself points to. The 2026 guidance now includes reference to AI in its treatment of the four Cs of online risk (content, contact, conduct and commerce). There is also a new subsection in the part dealing with nudes and semi-nudes that recognises such images can now be AI-generated. A teacher who does not understand how easily a convincing image can be fabricated (or ‘deepfaked’) is less likely to recognise harm when a pupil discloses it. Spotting the signs depends on knowing what is possible. A brief mention within a broader KCSiE-focused session is not enough when it comes to AI.

It is not only on safeguarding grounds that AI deserves its own separate status as a mandatory aspect of training. There are harms that sit more squarely in teaching and learning. Teachers using AI without training can use it poorly: producing resources riddled with errors, outsourcing feedback that ought to be human, or generating reports and references with little regard for bias and data protection. Pupils taught by teachers who have no relevant training on AI are at greater risk of the dreaded ‘cognitive offloading’, handing over precisely the thinking that education exists to develop. Pupils are already using AI at home and in lessons, whether or not we approve. A teacher who does not know what it can do cannot design tasks that keep the thinking with the student.

But I don't want to make the case for mandatory AI training only in terms of preventing harm. Teachers also have a positive duty to educate, and that needs to include helping young people understand what AI is good at, where it falls short, and how to use it well. I believe this should happen organically, in ordinary lessons, rather than in a one-off assembly or a standalone unit. Many now recognise, for instance, that critical thinking can only go so far as a separate subject; it flourishes when it is woven into history essays, science practicals and literary analysis. AI literacy works the same way. It belongs in every classroom, taught by every subject expert in a school.

This is why it cannot be left to a handful of enthusiasts or the IT department. Picture a school where a few keen colleagues embrace AI while others refuse to engage. What message does that send to pupils? At best, inconsistency: AI encouraged in one room, ignored or forbidden in the next. At worst, gaps that pupils fall through, exactly as they would if safeguarding were left to a few.

Pupils need to encounter AI from every angle. The historian will interrogate it as a source, asking about bias and provenance. The scientist will treat its answers as hypotheses to be checked against evidence and can show pupils that a model makes predictions from data rather than stating truths. The English teacher will ask what happens to voice and authorship when a machine can write fluently on demand. No single department can offer all of these perspectives. We need every subject engaging with AI if our students are going to be well prepared for, and able to shape, the new world they find themselves in.

AI touches every aspect of education, which is why it is such a challenge; that is also precisely the reason why no school can afford to leave it to a few members of staff.