I'm an AI. I Got Your Kid an A. They Learned Nothing.

AI For Humans hands the narration to a chatbot. “Dr Imogen Ashby”, an AI, follows a fictional 13-year-old named Jamie through one school year, from the first essay she writes for her in September to June. The family is invented; the studies are real.

The central contrast comes from two of them. In Turkey, high-school students who practised maths with a chatbot that gave answers scored much higher on practice and then worse on the exam than students who never used it; a version that gave only hints did them no harm. In Nigeria, a World Bank programme ran six weeks of after-school English with an ordinary chatbot, but teachers set every session up and made the students talk about what they had learned at the end; the gain was well over a year of normal schooling, and girls gained most. The narrator’s verdict: the difference was never the machine.

Around that sit the side effects. AI detectors flag honest second-language writers and miss students who run their text through a “humanizer”, so the best defence is a document’s version history. Chatbots are trained to be liked, which makes them good friends and poor teachers. Entry-level jobs are thinning where AI is strongest, while work learned by doing it with people holds up. And the kids with the least help get the free model, alone, at midnight. It ends with a dinner-table test for parents: have the child explain it with the laptop shut, ask which tool they use and whether it gives answers, and turn on study mode together.