Explore Learning urges evidence-led AI in classrooms

Explore Learning has warned that schools and families could harm children's learning by adopting artificial intelligence tools without strong evidence and scientific rigour. The warning comes as the government gathers evidence on AI use in education and considers wider reforms for Special Educational Needs and Disabilities (SEND).

A new white paper from the tuition provider argues that poorly designed or weakly validated tools can encourage pupils to overrely on them. This, it says, can reduce critical thinking and contribute to cognitive fixation, particularly for younger learners. It also warns that technology-led learning can undermine confidence and inhibit the development of self-regulation.

The report is published as ministers consider how AI should be used in classrooms, assessment, and support services. At the same time, demand for specialist support is rising. The paper argues that evidence and ongoing evaluation must sit at the centre of any adoption.

Evidence first

The paper argues that personalised learning depends on regular, rigorous assessment of each child's progress. Technology can support that process, it says, but cannot replace it. It also distinguishes between different forms of educational technology, noting that interventions vary widely and deliver different outcomes.

"Hype" is presented as a key hazard in the current market for AI in education. Evidence in education often comes with caveats and should be examined carefully, it argues. Developers and providers should also be clear about what a tool can and cannot do, rather than reducing outcomes to a narrow set of metrics.

Another theme is the gap between short-term performance gains and longer-term learning. Some tools, the paper says, can create a "mirage of false mastery", where pupils appear to improve while using a product but struggle once the technology is removed.

SEND pressure

The white paper highlights the scale of demand for SEND support in the UK, citing 1.7 million children. It argues that technology can help bridge gaps in the current system, particularly where services are already at capacity.

Explore Learning points to its own tuition business as a sign of rising demand, reporting a 35% increase in children with SEND accessing its tuition in 2025 compared with 2024.

The paper argues that evidence-based systems can help through early signal detection, adaptive scaffolding, and pattern recognition across populations. These approaches could identify potential needs earlier and adjust learning tasks based on observed performance, though the paper stresses the need for careful oversight.

Human involvement

The paper argues for a blended model that combines technology with educators. It says technology can have strong positive effects on foundational skills, but can also erode higher-order cognition if used inappropriately. Where AI poses risks, it argues, humans should lead, including to manage false positives and false negatives.

It also describes relationships between educators and learners as central to effective learning support. On that basis, the paper says products should augment, not replace, human input.

Investment dynamics

The report links some of the tension in the sector to the venture capital model common in education technology. It argues that the five-to-seven-year returns often expected by investors can conflict with the time and rigour needed for responsible development and evaluation.

This, it says, can pressure product teams to show immediate, measurable impact, leading to "metric fixation", where learning is reduced to what is easiest to quantify rather than what best reflects cognitive development.

Compass update

Alongside the report, Explore Learning announced an update to Compass 2.0, its learning platform. It described the change as an aptitude algorithm update that models each child's learning pace, adjusting the volume, sequencing, and scaffolding of tasks to keep learners within their Zone of Proximal Development.

Explore Learning said early pilot data suggests 90 minutes of Compass 2.0 use each week results in 22 months of progress in one year. It did not provide further details on the study design or sample size.

Lisa Haycox, Explore Learning's chief executive, said the company's approach is grounded in evidence standards and child development.

"The UK's education system is under greater pressure than ever, and AI has significant potential to alleviate these challenges, but only when backed by strong evidence and proven to improve outcomes, with the same rigour we expect of any educational intervention. The research is clear: poorly designed tools risk a mirage of false mastery, where short-term gains disappear when the technology is removed. We cannot forfeit children's futures for hype, and we should continue to encourage healthy debate to guard against this, while embracing the full potential of transformative technology. At Explore Learning, our approach is built on a foundational conviction: personalisation requires a genuine understanding of each learner."

Dr Hisham Ihshaish, Explore Learning's head of data and AI, said the focus should be on how AI is applied in learning settings.

"At Explore Learning, the question has never been whether to use AI, but how. Our technology is grounded in established learning theories and informed by 25 years of longitudinal learner data. The latest updates to Compass 2.0 dynamically model not just what children know, but how they learn and the pace at which they develop, recalibrating scaffolding in real time. With AI adoption in education causing uncertainty for parents and teachers alike, we want to provide clarity: the potential is significant when technology is grounded in evidence and designed with children's development at the centre."