How to teach Gen Alpha: are they actually harder to teach?
"Students today are harder to teach."
You hear it in staffrooms and you hear it from centre directors. Gen Alpha are harder to engage, more easily distracted, less willing to persist with anything that feels repetitive. To a teacher holding thirty-five children and a marking pile, it does not feel like an opinion.
It is worth taking the claim apart before building anything on it. The evidence is thinner than the confidence with which it gets repeated, and the part that does hold up points somewhere more useful than "children have changed".
What the attention data actually says
Between 14 and 28 May 2024, the US National Center for Education Statistics surveyed 1,714 public schools through its School Pulse Panel. Asked what was getting in the way of learning that year, 26% of school leaders said a lack of focus or inattention from students was having a severe negative impact and a further 49% said moderate. At 75% combined it was the most commonly reported barrier of any on the list.
Two things about that number before anyone leans on it. It is American, and it is school leaders reporting a perception rather than anyone measuring attention. What it establishes is that the people running schools believe focus is their biggest problem. That is not nothing, and it is also not evidence that Gen Alpha can concentrate for less time than the cohort before them.
The more interesting version of the observation is the one every teacher has made themselves. A child who cannot stay with a thirty-minute worksheet will spend two hours failing at the same level of a game. The game is more entertaining, certainly. It is also doing four things the worksheet does not: adjusting difficulty to the player, responding immediately, making progress visible, and being clear about what to try next.
None of that is an argument for gamifying a lesson. Those four properties belong to feedback, not to entertainment, and they are what a game borrowed from good teaching in the first place.
Feedback arrives after the thinking has moved on
A child spends an afternoon on a composition. It is collected, taken home, marked between lessons and admin, and returned a fortnight later. The marking may be careful and correct. By the time it lands, the child has stopped thinking about it.
There is a second problem underneath the timing, which is that feedback is not one thing. Wisniewski, Zierer and Hattie's 2020 meta-analysis, The Power of Feedback Revisited, pooled 435 studies and more than 61,000 participants. The headline effect of feedback overall was medium, d = 0.48, and the authors are explicit that the average conceals more than it reveals.
Split by how much information the feedback carried, the range is wide. High-information feedback, which told the learner something about the task, the process and how to regulate their own work, came in at d = 0.99. Reinforcement and punishment, which carry almost no information beyond a verdict, came in at d = 0.24. Roughly four times the effect, from the same activity done differently.
So "give students more feedback" is the wrong instruction. A tick, a grade and "good effort" are feedback, and they sit at the bottom of that range.
Three children, one worksheet
In a class of thirty-five, one child has a skill secure, a second is missing the foundation underneath it, and a third understands it but cannot yet use it without being prompted. All three receive the same worksheet and the same homework.
Teachers know this. They can usually name the three children. What they are trying to do has a name, differentiated instruction, and the idea is decades old.
The constraint has never been insight. It is arithmetic. To personalise properly you need a per-skill picture of every child, material aimed at each gap, and a check that the gap closed. Across thirty-five children and multiple classes, the first is a marking workload nobody has, the second is a preparation workload nobody has, and the third almost never happens. Differentiation in practice collapses into three worksheets labelled by ability, set once and never re-checked.
Where the hours actually go
TALIS 2024 surveyed about 3,500 teachers and principals across all 145 public secondary schools in Singapore, plus ten private ones, between April and August 2024. Full-time teachers reported 47.3 hours a week against an OECD average of 41, with 6.4 of those hours on marking and correcting student work against an OECD average of 4.6, and about four hours on general administrative work.
The stress figures follow the hours: 53% named too much administrative work as a source of significant stress and 49% named too much marking.
One detail in that data cuts against the easy version of this argument, so it belongs here rather than in a footnote. Marking time in Singapore has gone down by about an hour a week since TALIS 2018, while total working hours went up. The growth is in lesson planning, counselling, co-curricular activities and communicating with parents. Marking is not a workload crisis getting worse. It is a steady 6.4 hours, 40% above the OECD average, that half of teachers name as a stressor, and it sits directly between a child doing something and that child finding out how it went.
That last part is why it is worth attacking, and it is a narrower claim than "teachers are drowning".
Shortening the distance
A teacher should not have to spend an evening reading thirty-five compositions to discover that eight children are missing the same thing.
What Zippy does with that evening is specific. Every piece is marked against the criteria you gave it, per criterion and per skill, with each comment tied to the sentence that earned it. Those skills accumulate into a record for each child across the term: secure here, developing there, missed in three of the last four pieces. From that record Zippy drafts practice aimed at what the marking actually found, for one child or for all eight carrying the same gap. You read it and change it before anything reaches anyone. It comes back marked against the same skills, so you can see whether the gap closed.
The fifth step is the one most personalised-learning products leave out. Generating differentiated worksheets is easy now. Checking whether they worked is what makes it learning rather than activity.
What this does not solve
An AI marker handed no rubric agrees with a human marker 33.5% of the time, measured by researchers at the University of Georgia (arXiv 2504.13557). Given the teacher's own rubric it clears 50%. Neither number is a marking tool you can leave alone, which is why nothing here is assigned to a child without a teacher reading it first.
A diagnosis is also not a lesson. Knowing that eight children are weak on paragraphing does not teach paragraphing to them. It tells you what Tuesday is for, which is a real improvement on not knowing and considerably less than the "AI tutor for every child" framing implies.
And the evidence in this post has holes worth naming. The attention data is American and secondary-weighted. The workload data is Singapore secondary teachers only, so the primary classroom where PSLE writing actually happens is not directly measured by it. Nothing here shows that faster feedback improves outcomes for Gen Alpha specifically, because that study has not been done.
What is left after the caveats is still enough to act on. Feedback carrying real information works about four times as well as a verdict, teachers spend 6.4 hours a week producing it, and most of it arrives after the child has moved on. Closing that gap does not require anyone to decide whether a generation has changed.
Frequently asked questions
Are Gen Alpha students harder to teach?
The strongest available evidence is perception rather than measurement. In the May 2024 NCES School Pulse Panel, 75% of US school leaders said student inattention was having a moderate or severe negative impact on learning, the top barrier reported. No study establishes that Gen Alpha can concentrate for less time than earlier cohorts.
How does Gen Alpha learn differently?
They are used to systems that respond immediately, adjust to them and show progress. Classroom learning that relies on delayed feedback and one worksheet for everyone offers none of those three, which changes what feels engaging rather than what is possible.
How can teachers engage Gen Alpha students?
Shorten the gap between a child doing something and finding out how it went, and make the feedback carry information rather than a verdict. Feedback about the task, the process and self-regulation measured d = 0.99 across 435 studies, against d = 0.24 for reinforcement and punishment.
How much time do teachers spend marking?
Full-time secondary teachers in Singapore reported 6.4 hours a week on marking and correcting student work in TALIS 2024, against an OECD average of 4.6, within a 47.3-hour week. 49% named too much marking as a source of significant stress.
How can AI help teachers personalise learning?
By carrying the arithmetic, not the judgement. AI can mark against a teacher's rubric, build a per-skill record for each child and draft practice aimed at the gaps it found. An AI marker given no rubric agrees with a human only 33.5% of the time, so the teacher reviewing and approving the output is the mechanism, not a disclaimer.
Sources: student inattention, NCES School Pulse Panel, May 2024. Feedback effect sizes, Wisniewski, Zierer & Hattie, Frontiers in Psychology, 2020. Teacher hours and stress, TALIS 2024 via OECD Education GPS. AI/human marker agreement, arXiv 2504.13557.
If you want the argument underneath the personalisation claim, personalised learning at class scale takes apart the two sigma study most AI education pitches are built on. The marking side is covered in how teachers stay in control of AI marking.