Students are already ahead — and that is the problem
Picture a Year 10 student in Grantham sitting down to draft a history essay. Before opening a textbook, she opens ChatGPT. She has been doing this for months — refining prompts, checking outputs, occasionally disagreeing with what it produces. Her school, like most in England, has no explicit AI policy that names the tool she is using.
This is not a story about a rogue pupil. By 2024, 77.1% of UK teenagers aged 13–18 were using generative AI routinely — nearly double the figure from the year before, according to the National Literacy Trust's Annual Literacy Survey of 53,169 pupils. Grantham's secondary students are part of that shift. The tools spread faster than any formal response could follow.
The instinct is to shrug and call it progress: young people are digital natives, they adapt, they figure things out. BCS, the Chartered Institute for IT, argues that assumption is precisely the problem. Using an AI tool and understanding what it is doing are not the same skill. A student who can prompt ChatGPT fluently may have no grasp of how the model was trained, what biases it may carry, or why it sometimes confidently produces nonsense. Confidence with the interface is not the same as literacy about the technology.
The gap Grantham's schools face, then, is not one of devices or connectivity. It is a gap between adoption and comprehension — and at present, the students are well ahead.
What individual schools are doing — and why it varies so much
The variation starts at the top of the governance chain. When the DfE updated its AI guidance in August 2025, it did so deliberately without mandating a specific curriculum module. Schools and colleges set their own rules on AI use; final responsibility sits with the head teacher and governing body. That design creates genuine flexibility — but it also means the floor is whatever each school decides to build.
In Grantham, the result is a marked spread. The King's School uses the Century AI adaptive learning platform to personalise pupil learning, and its GCSE and A-Level Computer Science syllabi explicitly cover the ethical, cultural, and legal consequences of automation and AI. Kesteven and Grantham Girls' School (KGGS) takes a different but equally deliberate approach: AI's societal impact is woven through both its computing curriculum and pastoral guidance sessions, with particular attention to the gender gap in technology careers. Both schools have made active, documented choices about what AI education means for their pupils.
Where schools have not made that choice centrally — and where structured provision is less fully recorded — the picture becomes harder to map. Rather than single out individual institutions, it is more honest to note that gaps in the documentary record reflect a system that offers no guaranteed baseline: if a school does not choose to build AI provision, nothing requires it to.
The voluntary programmes filling part of this gap do genuine good where they land. The Raspberry Pi Foundation's Experience AI — co-developed with Google DeepMind and available free to KS3 teachers — reports that 93% of participating teachers gain increased conceptual understanding of AI, and 87% feel more confident teaching it. But participation is optional. A school whose leadership has not prioritised AI professional development is unlikely to opt in. That one word — opt-in — is where the unlevel system shows most clearly.
The reforms that will help — but not yet
England has, at least, acknowledged the structural problem. In November 2025, the government confirmed it would replace the existing, narrowly focused Computer Science GCSE with a broader Computing GCSE — one that integrates AI literacy, data science, media literacy, and digital ethics alongside traditional coding. It is the most substantive change to this area of the national curriculum England has made, and it is genuinely forward-looking rather than cosmetic.
The difficulty is the calendar. Mandatory teaching under the new framework begins September 2028. The final reformed curriculum is not even scheduled for publication until spring 2027. Any student currently in Years 7 to 10 will have sat their GCSEs and left secondary school before a single lesson of the reformed course is taught. The reform is real; its benefits are not available to the cohort in Grantham's classrooms today.
A proposed Level 3 qualification in data science and AI — an A-level equivalent for 16–18-year-olds — adds a meaningful post-16 pathway and signals that policymakers understand the depth of change required. But with the curriculum still under development and publication not due until 2027, students making sixth-form choices right now are doing so without a settled framework in place.
None of this is a reason to dismiss the reforms. The DfE's November 2025 announcement represents a clear institutional recognition that the previous settlement — AI tucked inside an optional GCSE taken by a minority — was no longer adequate. The problem is that recognition and remedy operate on different timescales. For the students already navigating an automated economy, the reforms are a promise that arrives after they have already left.
The economy students are walking into
Grantham's manufacturing economy is not waiting for the curriculum to catch up. Food processing plants, precision engineering firms, and energy supply chain businesses across the town are already operating with automated systems that require digital interface literacy from workers at every level — not just from engineers.
The employers are named and local. Bakkavor, one of the largest food manufacturers in the region, runs plants in which automation is integral to production. Siemens Industrial Turbomachinery, significant across the South Kesteven supply chain, expects digital competence well below senior engineering grades. Pentangle Engineering Services and Olympus Automation — both active in the area — recruit around workers who can interface with control systems and interpret automated processes. Controls engineers and PLC programmers in this sector typically earn £35,000 to over £60,000. But the salary range is not just evidence that specialist roles pay well; it signals that this is mainstream demand, not a niche corner of the labour market. Even operative-level positions increasingly carry a digital literacy requirement that did not exist a decade ago.
Greater Lincolnshire's Local Skills Improvement Plan, launched in January 2026 by the FSB in partnership with the Greater Lincolnshire Combined County Authority for Skills England, formalises what individual employers have been saying informally. Spanning 41,000 businesses and a combined workforce of 550,000, the LSIP names Industry 4.0 alignment as a central priority. It is a regional institution committing to paper that automation in manufacturing is moving faster than the training structures built to support it — a concession worth noting because institutions of that kind rarely frame a problem until it has become impossible to sidestep. For students leaving Grantham's secondary schools today, that is not a warning about the future. It is a description of the economy they enter the morning after results day.
The sharpest gap: leaving school at 16
Good post-16 provision exists in Grantham. The College's Institute of Technology campus offers dedicated AI modules, engineering robotics, digital bootcamps, and advanced embedded systems courses, with employer partnerships that include Siemens Industrial Turbomachinery and Olympus Automation. Students who progress there into apprenticeships, NVQs, or short specialist programmes are stepping into genuinely practical preparation for the automated economy described in the previous section.
The structural problem is simpler than it might appear: that provision only reaches the young people who walk through the door.
For those who leave education at 16 — entering the labour market that same morning after results day — the available AI literacy is whatever their secondary school delivered. And the evidence suggests that delivery has been uneven. Machine learning concepts and digital ethics appear mainly within Computer Science lessons or voluntary opt-in programmes. Computer Science is a minority GCSE choice; most students do not take it. That means the majority of school leavers have had no consistent, structured route to AI understanding — not because schools have failed in any simple sense, but because no mandated baseline existed to guarantee one.
BCS is direct on this point: the assumption that young people are automatically 'digitally native' does not hold. Using AI tools and understanding them are different things, and the latter requires deliberate teaching. For the 16-year-old heading into a food processing plant or a component workshop, neither the reformed Computing GCSE nor the Level 3 data science qualification arrives in time. What they carry with them is whatever happened — or did not happen — in the years before.
What better AI teaching actually looks like
A 2023 systematic review of 50 AI education studies found that the most effective secondary AI pedagogy involves collaborative, project-based learning across disciplines — students building things together, solving problems that cut across subjects, and encountering AI in genuine context. Younger secondary students learn best by experiencing it directly; older students can engage with more technical components. Neither approach fits naturally inside a single optional GCSE.
Computing education researchers are consistent on the structural implication: AI is an interdisciplinary field, drawing on mathematics, social sciences, philosophy, and linguistics. Teaching it only where a student has opted into Computer Science is likely to miss most of the school population. Teaching it through geography, science, history, and PSHE — embedding it where the majority of students already sit — is what the evidence favours. But that requires deliberate curriculum design and the teacher confidence to deliver it.
That is where the actual gap lies. Not in the availability of technology — schools have screens, students have devices — but in the architecture of what is taught, by whom, and to which students. The reformed Computing GCSE will eventually move the baseline. Teacher development programmes can shift classroom confidence. But the structural and pedagogical changes needed take time that the current cohort does not have. For students leaving Grantham's secondary schools before September 2028, the question is not what AI education will eventually look like. It is what they leave with now — and whether it is enough for the automated economy waiting on the other side of results day.
- [1] Artificial intelligence (AI) literacy education in secondary schools: a review. (2023). https://doi.org/10.1080/10494820.2023.2255228 https://doi.org/10.1080/10494820.2023.2255228