There are endless conversations about how artificial intelligence might make educators more productive by reducing administrative burdens. Yet for all the discussion, solid evidence has been scarce: plenty of anecdotes, but little systematic research. That is beginning to change.
The Department for Education recently asked Ofsted to investigate how “early adopter” schools and FE colleges are embedding AI to manage risks, support teaching and learning, and streamline admin. Ofsted’s report concludes that AI can indeed save teachers time, and the evidence of its impact on teaching and learning is steadily growing.
Meanwhile, the Education Endowment Foundation (EEF) is recruiting primary schools to test Oak’s AI-powered lesson planning tool. The tool promises to cut planning time by 3 – 4 hours per week by generating lesson plans, adapting resources, and suggesting activities for different ability levels. Results are due in autumn 2026, adding to the global evidence base. This trial follows an earlier EEF study showing that teachers using ChatGPT reduced lesson planning time by 31% — though researchers cautioned that more work is needed to understand how AI can support teachers without compromising quality.
Beyond education, economists often distinguish between two roles AI can play:
● Automation: AI fully carries out a task without human involvement. ● Augmentation: AI supports humans, requiring oversight and complementing their work.
Anthropic, the company behind Claude, analysed a million text-based conversations and found that educators and librarians score highest for augmentation: around 40% of their job tasks can be enhanced by AI. In other words, teachers are unlikely to be replaced, but many are already benefiting.
Survey data backs this up. A recent Teacher Tapp poll found that a quarter of teachers had used AI for school work in the past week, more than half had tried it at least once, and one in ten secondary teachers had used it during a lesson. English and primary school teachers are the group most likely to have used AI in the past week – not the maths and computing teachers which people sometimes expect!
But is it really saving time? The gains so far are modest. Even the keenest adopters often describe scenarios like: a resource that once took two hours to write now takes ten minutes. A modest augmentation to an existing process. Helpful, but hardly revolutionary. In my 20 years of teaching, I’ve seen many “edtech revolutions” fizzle into small efficiencies rather than systemic change. Sometimes they even create more work. Is AI destined to follow the same path — giving us better emails, slicker worksheets, and incremental improvements rather than transforming the sector?
For now, I’ll take the small wins. Time saved on routine tasks is still valuable. If AI helps teachers spend more energy on building relationships with pupils, designing creative lessons, and supporting diverse learning needs, that is progress worth having.
