England Adds AI Literacy to the Curriculum. Only 14% of Teachers Are Ready
England's curriculum reform makes AI literacy compulsory and replaces the Computer Science GCSE, but a Royal Society survey found just 14% of teachers feel confident teaching it.
A curriculum promise with a 2028 horizon
On 5 November 2025, the Department for Education accepted the core recommendations of the independent Curriculum and Assessment Review, published as Building a World-Class Curriculum for All. Among the headline changes: the narrowly focused Computer Science GCSE is being replaced with a broader Computing GCSE, and a reformed national curriculum will make AI literacy, digital literacy and media literacy compulsory content across key stages, with implementation expected from 2028.
The review's reasoning is blunt. It rejects the assumption that young people absorb digital literacy simply by growing up online, and argues that pupils need to understand, from an early age, how computers and AI systems are trained on data, not just how to use the apps built on top of them.
What the GCSE rewrite actually changes
The new Computing GCSE is meant to cover "the full breadth of the Computing curriculum" rather than the narrower programming-and-theory focus of the current specification, while still allowing progression to A-level Computer Science and specialist computing careers. The government is also exploring a new Level 3 qualification in data science and AI for 16-to-18-year-olds, and has committed to continued funding for the National Centre for Computing Education, which trains and supports computing teachers.
Julia Adamson MBE, executive director of education and public benefit at BCS, the Chartered Institute for IT, welcomed the changes, describing digital skills as "essential for full participation in modern society" and calling the new qualification a genuinely "future-facing computing GCSE."
That is the policy. The workforce meant to deliver it is a different story.
The readiness gap
In March 2026, the Royal Society ran a weighted survey of more than 9,000 teachers in England, submitted as written evidence to the Education Select Committee. It measured confidence across three dimensions of AI literacy: technical understanding, practical tool use, and the ethical and human dimension of AI in society.
Only 14 per cent of teachers surveyed said they felt confident across all three dimensions of AI literacy. A third reported confidence limited to practical tool use alone, and a third reported limited confidence across the board.
The same survey found that just 18 per cent of respondents had received any structured professional development on AI, 37 per cent said AI literacy was not currently addressed anywhere in their school's teaching, and only 2 per cent said it was covered across most subjects rather than confined to computing lessons.
That last figure matters most for a policy that is meant to sit across the curriculum, not inside one department. If AI literacy is being written into geography, English and citizenship as much as computing, the current picture suggests almost none of that cross-curricular delivery exists yet.
Why Ofsted's silence compounds the problem
A structural reason the gap may persist: computing is currently underrepresented in Ofsted's evaluation framework. Because inspectors rarely probe it in any depth, schools have limited incentive to prioritise investment or timetable space for it over subjects that are visibly inspected. Sector commentary accompanying the curriculum reforms has argued that AI literacy will not be valued or delivered consistently until it becomes visible in the inspection framework itself, otherwise a compulsory curriculum requirement risks becoming a compliance formality rather than a taught reality.
The charity moving faster than government
Some of the running on AI literacy is being made by the third sector rather than Whitehall. The micro:bit Educational Foundation, a UK charity best known for the pocket-sized coding device used in British primary schools, has spent 2026 expanding a CreateAI programme designed to help pupils become creators of AI systems rather than passive consumers of them. Its most developed pilot, however, launched this year with 300 educators across the United States, backed by Google.org, rather than in the UK schools the Foundation was originally built to serve. It is a telling detail: a British charity's flagship AI-literacy effort is scaling abroad while England's own statutory version waits for a 2028 rollout and a still-pending public consultation on the detail of qualifications.
What to do
- Do not wait for 2028. Audit what AI literacy content already exists across departments, not just in computing, and identify the gaps now rather than at the point of statutory rollout.
- Treat the Royal Society's three dimensions, technical understanding, practical use, and ethical and human context, as a planning checklist for CPD, since most current teacher confidence sits in practical use alone.
- Push AI and digital literacy into school self-evaluation and improvement plans ahead of any Ofsted framework change, rather than waiting for inspectors to ask.
- Watch the National Centre for Computing Education for updated training offers tied to the reformed GCSE and the proposed Level 3 data science and AI qualification.
What to watch
- The DfE's public consultation on the reformed curriculum and qualifications, expected later in 2026, which will fix the detail of what the new Computing GCSE actually assesses.
- Whether Ofsted's next framework revision gives computing and AI literacy the visibility that sector bodies say is currently missing.
- Take-up and findings from the micro:bit Foundation's CreateAI pilot, as an early indicator of what works before England's own statutory content is finalised.
- Further Royal Society or DfE data on teacher confidence, to see whether the 14 per cent figure moves before the 2028 implementation date arrives.