Labour prime minister Keir Starmer promised his government would bring about political “change.” Yet his government appears just as excited by the potential for artificial intelligence (AI) to transform education as its predecessor, which is quite an achievement. It is even using the same language about AI—perhaps artificially generated in both cases: “turbocharging” learning.
Science secretary Peter Kyle said yesterday that it was “OK” for children to use ChatGPT—an AI tool which can generate words and images in response to user prompts—to do their homework. He raised the usual caveat that this application would depend on “supervision,” ignoring—again, as usual—the ability of tech-savvy youngsters to find ways around whatever barriers their less-aware elders put up.
Kyle also laughed off the potential harms of AI misuse and told his interviewer, in an assuredly patronising tone, that “I’m of an age where we had this conversation about calculators.”
The comparison is weak, not only because of our palpably problematic relationship with calculators (who now does not often turn to one even to confirm the simplest of sums?) but also because of the much broader scope of AI’s potential uses. Pupils across the country are, for example, already using the ever-developing technology to produce indecent images of their classmates.
Arabella Skinner, the children’s campaigning director of UsForThem’s SafeScreens Campaign, told europeanconservative.com it is “incredible that once again we are hurtling straight into something we do not fully understand, have no risk analysis for, nor any evidence as to whether it will help rather than harm our children.”
AI is currently full of mistakes and has the potential to change how young brains develop. The safeguarding risk of the content that children could be exposed to alone should be enough to stop this in its tracks.
Kyle’s approach is effectively that of a defeatist—that children should be encouraged to replace the old (and tested) with the new “because ChatGPT and the AI technology that’s using language is already being used across the economy.” The idea that his department could test the impacts of new such technologies on children convincingly—and, for that matter, on adults—and even attempt to turn the tide if it found these were bad does not seem to have registered. Besides, turning back would not fit in well with Starmer’s vision of Britain as a “great AI superpower.”
So long as AI doesn’t become a “replacement for the expertise of teaching”—presumably just in particular undefined areas, since it already has in some—Kyle and company will be happy.
