AI and the age gap in formative assessment

The gist. A survey of 60 English-language teachers in Kazakhstan found that younger teachers reported using every formative-assessment practice, including digital and AI tools, far more than colleagues aged 46 and over. The researchers argue that AI could help close that gap, but their own study never tested whether it does. Read this as an early signal about a generational habit gap, not proof that AI is the fix.

What the study found

The team surveyed 60 teachers of English across public schools in Almaty city and region, split evenly into three age bands (24-34, 35-45 and 46 and over, with 20 in each) and spanning 2 to 25 years of experience, then interviewed 12 of them in more depth. The study named chi-square tests and ANOVA in its methods but reported no test statistic or p-value for any age comparison, so the differences below are self-reported percentages rather than confirmed statistical effects.

The pattern was consistent: younger teachers reported more of every practice. On using digital or AI tools to support assessment, 40% of the 24-34 group said they did, against 15% of the 46-plus group. For quick in-class checks the split was 90% to 50%, for peer-assessment 55% to 20%, for regular self-assessment 50% to 30%, and for open-ended questioning 75% to 50%. Teachers pointed to time constraints, large classes, limited training and low learner motivation as the main barriers, and the AI examples in the paper, such as a quick poll that revealed a class had misunderstood possessive apostrophes, were single teachers' stories rather than measured results.

What it might mean for your classroom

These are teachers in Almaty's English-language classrooms, so treat the numbers as a prompt to look at your own department rather than a UK benchmark. The useful question is what younger colleagues are doing that older ones are not.

• Map the gap in your team. Notice who already leans on quick checks and digital tools, and pair them with colleagues who prefer more traditional, teacher-led approaches.
• Try one lightweight check. A single digital tool for an instant class poll can surface a misconception in seconds, the mechanism these teachers valued most.
• Pilot before you scale. The paper's own authors recommend small pilots precisely because the effect of AI has not been tested here, so gather your own evidence before committing a department.

Treat it as an experiment, not a recipe: pick one formative-assessment routine, add a simple tool, and check whether it genuinely tells you more about your learners than your current approach does.

How solid is this?

Early signal. The age gap is real in the sense that it appears across every practice the survey measured, but two things hold the grade down. First, the study names chi-square and ANOVA yet reports no test statistic or p-value, so the gap is a descriptive difference in what teachers said, not a confirmed significant effect. Second, and more important, the headline idea that AI can close the gap is never tested: the authors state plainly that "without experimental testing of AI's direct impact on learning, the suggested approaches still need further validation", and their AI examples are one-teacher anecdotes rather than data from the 60-teacher sample. The sample is also limited to Almaty and may not represent rural or non-Russian-speaking Kazakhstan, and all of it is self-reported. A controlled trial of an AI tool against a comparison group would raise the grade.

The source

Yersultanova, G., Orazakynkyzy, F., Zhyltyrova, Zh., & He, Ling. (2026). Integrating AI into formative assessment in English language classrooms of Kazakhstan's schools. Bulletin of Zhetysu University named after I. Zhansugurov, 2(119), 68-80.

Read the original study