Imagine glancing into the near future of education. In Sweden, Lexplore uses AI-powered eye-tracking not just to assess reading, but to actively screen children for potential difficulties like dyslexia, identifying needs often before they become entrenched problems. What’s striking isn’t just the technology; it’s the philosophy. This feels less like conventional testing or inspection, and more akin to a proactive health service identifying risks and enabling early support. Could this preventative, diagnostic approach be a model for England’s Department for Education?

The UK’s National Health Service operates on a foundational principle: catch health issues early through reliable and inexpensive screening. Interventions triggered by timely screening are invariably more personalised, cheaper, more successful, and less intrusive. Think about developmental checks for babies or targeted screening programmes for adults based on age and risk factors – it’s not a one-size-fits-all annual exam for everyone simultaneously. What if we applied this proven, preventative philosophy to education, but converged with the power of AI?

We stand at the cusp of an explosion in convergence between artificial intelligence and educational technology. AI is rapidly moving beyond simple automation; it can increasingly ‘observe’ and analyse the process of learning. Imagine systems that don’t just mark a final answer, but watch how a pupil reads – tracking eye movements for fluency and comprehension indicators. Picture AI analysing handwriting formation for potential motor control issues or even just letter formation; identifying patterns in mathematical problem-solving that might suggest dyscalculia or misconceptions; observing collaborative interactions in virtual environments; or even assessing biomechanical efficiency in PE. While acknowledging the critical ethical considerations around data privacy and algorithmic bias that must be addressed, the potential for deep, nuanced understanding of individual learning is immense. And it will arrive. Don’t think the tidal wave of AI is going to miss education. It’s going to cover every single bit of it.

This capability of AI allows us to envision a shift away from assessing children primarily to measure school performance, towards screening individuals to understand their specific needs. Contrast this with our current reliance on blunt, mass-approach strategies. Pupil Premium funding, while well-intentioned, often lacks the granular data to target underlying needs effectively. Large-scale EEF randomised controlled trials dictate averaged-out ‘best practices’ that may not suit every child or context. Rigid, centrally mandated phonics schemes meet pupils at varying developmental stages.

Consider the annual phonics screening check – the infamous graph plotting average scores by birth month, a near-perfect downward slope from September to August-born children, is a stark illustration. It highlights the absurdity of assessing every child at the same chronological point, ignoring months of developmental difference. The check itself may have value, but the one-size-fits-all process is flawed. It’s a system designed for cohort-level data collection, not individual diagnosis. Similarly, high-stakes standardised tests often narrow the curriculum, induce stress, and provide only a snapshot in time, failing to capture progress, promote creativity, or cultivate critical thinking.

Imagine, instead, dynamic, AI-powered screening. Phonics checks could be triggered by birth month, not school year cohort. Algorithms could identify children needing earlier or more frequent screening based on a growing profile of risk factors – perhaps language delay, family history, or early indicators from those AI observations of reading or writing. A five-year-old wearing an eye patch for 18 months wouldn’t just potentially ‘fail’ a single test; their progress could be sensitively tracked via regular screening against national benchmarks for learners with similar challenges.

The data generated wouldn’t primarily serve to rank schools, a practice often misleading given the vast differences in intake, funding, and context. Instead, it would empower precise, personalised interventions. AI analysis identifying specific phoneme difficulties could trigger targeted support from a school’s in-house reading specialist. Real-time assessment of maths understanding could dynamically adjust adaptive learning software. Observed motor control difficulties could lead to specific occupational therapy recommendations. This approach allows resources – human expertise, tailored software, specific aids – to be channelled effectively, supporting a child directly.

Scaling this vision creates a powerful national dataset focused on children’s learning needs and progression trajectories, not crude school comparisons. This brings us back to the idea of a “National Education Service.” While the term was politically championed by Labour in recent years with a focus on universal access and lifelong learning, this technologically-enabled vision offers a different emphasis: a service philosophy built on proactive, individualised screening and support. It uses AI not for judgment, but for deep understanding, enabling interventions that are early, cost-effective, successful, and minimally intrusive where possible.

Isn’t it time the DfE seriously considered shifting its focus from ranking schools through mass assessment to truly nurturing every child’s potential through intelligent, personalised screening? Perhaps a reimagined NES, powered by ethical AI, is the future. It’s already happening at an elite sport level, so why not be bold and have a plan to use it for every child in the country.

Dr James Shea @englishspecial

Image of elite sport using AI to ‘watch’ a player’s perfomance