What does the national curriculum say about AI - and what are schools supposed to do about it?

The curriculum hasn't kept up

The national computing curriculum was introduced in 2014. It was a genuine step forward at the time - replacing ICT lessons with something more substantive, teaching children how computers actually work rather than just how to use a spreadsheet.

But 2014 was before the smartphone became ubiquitous in children's lives. Before social media platforms were designed around algorithms that shaped what young people saw and believed. And long before generative AI tools like ChatGPT put a remarkably convincing, remarkably unreliable information engine in the hands of every child with a phone.

The current national curriculum makes no mention of artificial intelligence. Not in the computing programme of study. Not in PSHE. Not anywhere. Children are using AI tools daily - often without understanding what they are, how they work, or why that matters - and the statutory framework schools operate within has nothing to say about it.

Change is coming - but not yet

The government knows there's a problem. In November 2025, following an independent Curriculum and Assessment Review, it announced plans to reform the computing curriculum. The reformed version will introduce age-appropriate AI concepts from KS2 onwards - helping children understand what AI is, how it makes decisions, where it already exists in their lives, and how to use it responsibly.

A new, broader Computing GCSE will replace the existing qualification, bringing AI and data skills into secondary education for the first time as statutory requirements. A post-16 qualification in Data Science and AI is also being explored.

What the government is - and isn't - asking schools to do

Faced with a curriculum that hasn't caught up, the Department for Education has taken a few interim steps.

In June 2025, the DfE rolled out free teacher training materials covering AI basics, using generative AI safely, and developing school AI policies - created in partnership with Chiltern Learning Trust and the Chartered College of Teaching. Schools are encouraged, but not required, to use them.

The updated RSHE statutory guidance, due for implementation in 2026, explicitly names AI-generated deepfakes as a safeguarding concern - making it one of the first statutory references to AI in the school context. This creates a specific duty for schools around online safety and synthetic media, even if the broader curriculum hasn't moved yet.

Schools are also expected to have AI policies in place by the end of 2026, covering how AI tools are used by staff and pupils, and how personal data is handled when AI is involved.

But none of this amounts to a requirement to teach children what AI actually is - how it works, where it exists in their lives, and what it means for their futures. That part remains entirely voluntary.

Why primary schools are in the hardest position

Secondary schools can, in theory, bolt AI content onto existing computing lessons or fold it into PSHE. Primary schools face a steeper challenge.

A 2024 survey by the National Centre for Computing Education found that 40% of primary teachers rated their confidence teaching computing as low or very low. That's the baseline - before AI is added to the picture. Adding AI concepts to the curriculum without significant support risks widening an existing gap rather than closing it.

Primary teachers are generalists. They are brilliant at many things, but they were not trained to explain how a large language model works, or why an AI might produce a convincing but entirely false answer, or what a deepfake is and how a child can recognise one. That's not a criticism - it's simply not what the system asked of them. Until now, it didn't need to.

What children are doing in the meantime

While the policy catches up, children are not standing still. Research published in late 2024 found that nearly four in ten secondary school pupils were already using generative AI tools outside school, with no structured guidance on how to use them critically or safely. By the end of that year, that figure had risen to over 70%.

Primary age children are not far behind. The same tools are available to them - and in many cases, actively used - whether or not schools address it. The question facing head teachers is not whether their pupils will encounter AI. They already have. The question is whether those children have any framework for understanding it.

What schools can do now

The honest answer is that there is no off-the-shelf statutory solution. Schools that want to address the AI literacy gap before 2028 are doing so through a combination of:

• Whole-school assemblies that introduce AI concepts to entire year groups at once
• Integrating AI awareness into existing PSHE and computing lessons where teacher confidence allows
• Bringing in external specialists to deliver sessions that staff don't yet feel equipped to lead themselves
• Using the new DfE teacher training materials to build staff confidence over time

The RSHE safeguarding update gives schools a concrete compliance hook for the deepfakes and online safety piece specifically. For many school leaders, that's the most urgent entry point - it's a statutory requirement, not an optional enrichment activity.

The broader question of AI literacy - what AI is, how it works, what it means for the world your pupils are entering - is still optional. But the schools doing it now are building a foundation that will matter increasingly as their pupils move into secondary education and beyond.

The curriculum gap is not schools' fault. It's a structural lag - the inevitable result of policy moving slower than technology. But children don't get to wait for policy to catch up. They are navigating an AI-shaped world right now, with or without the tools to do it well.

The question for school leaders isn't whether AI literacy matters. It's whether your pupils can afford to wait until 2028 to start building it.