Pupils struggle to spot untrue AI content – report (www.bbc.com)

An Oxford University Press survey of more than 2,000 pupils aged 13–18 finds widespread classroom use of AI but significant difficulty in judging its truthfulness. Eight in 10 teenagers said they use AI for schoolwork, yet just over half reported it’s hard to tell whether AI-generated content is true; nearly a third said they couldn’t tell and a further fifth were unsure. While students credited AI with boosting skills like problem‑solving, creative writing and critical thinking (nine in 10 said they’d benefited), many rely on teachers for help—yet about a third feel their tutors lack confidence using AI tools. Educators report mixed behavior: some pupils use AI effectively as a tutor, others copy‑paste unchecked answers, and even adults can struggle to spot synthetic people in video. The findings matter because they highlight a gap between rapid student adoption and limited verification skills, exposing schools to risks from hallucinations and misinformation inherent in current generative models. The report underlines the need for targeted teacher training, curriculum changes to teach source evaluation and prompt literacy, and better school policies and tools to detect and contextualize synthetic content. OUP has launched an AI and Education site and the UK Department for Education published guidance—steps toward equipping teachers and students to use AI safely and critically.