Just after the October half-term break, the education secretary, Bridget Phillipson, pronounced to an audience at the Confederation of School Trusts conference that her “challenge” to school leaders and trusts was to become more inclusive - for every child to be achieving and thriving.

It was something of a surprise, then, to be told later in November that one of the first steps Phillipson intended to take was introducing new targets for attendance.

Not just any old targets, mind you - oh no: these are AI-generated, apparently.

School attendance targets

No teacher in any school thinks that poor attendance is OK. I can’t imagine there’s a school in the land that doesn’t work to try to improve attendance.

But a new target won’t magically make that any easier. And it might actually do some damage to the attempts to be more inclusive.

Any headteacher might think twice now about agreeing to support a pupil with appropriate alternative provision, if that support might also detract from work elsewhere to improve attendance.

Is it better to have a pupil attending part-time and accessing learning tailored to their need, or to insist on upping the number of “present” marks? Can the magical artificial intelligence account for that?

Health impacts on education

Will the AI cauldron produce figures that reflect the challenges of pupils with significant medical needs who find themselves at frequent appointments in an NHS that struggles to meet basic need, let alone accommodate school timetables?

Has a carefully-worded prompt considered the fact that some schools hold children on their books whom they know will never attend, because they’re awaiting appropriate specialist provision that local authorities struggle to provide?

No headteacher should ever have to consider the impact on the figures of accepting a pupil on to the school’s roll - but a new target will inevitably make it a consideration.

School absence improvements

The job of raising attendance is already hard enough. Too many attempts to engage with families struggling to maintain good patterns of attendance already result in parents complaining that schools only care about data.

Even when schools are working their hardest for the benefit of pupils, it’s harder to shake that perception of being data-driven if the secretary of state decides to drive things with data.

And there’s no hiding place for the department to claim that schools won’t be held accountable for these new targets. The messaging is already clear: fail to meet your target next year and you can be added to the list for additional support - whether you want it or not.

Comparing schools

In the meantime, you can contact a school that already has higher attendance than you, despite the AI-crunched similarities.

What advice might they give? Is it raffle prizes for the punctual? Minibuses to round up the ruffians? Or maybe just getting lucky with families who are keen for their children to attend and learn?

Presumably we are all supposed to be so dazzled by the mere mention of AI that we should just accept its wisdom and strive for its goals. But if there’s one thing teachers can probably tell you more than most, it’s that AI is not a panacea. Worse: it’s something of a black box.

The very nature of the artificial generation of content is that we’re kept in the dark when it comes to how figures are arrived at. Rather than careful analysis of numerical patterns combined with expert human knowledge, we’re instead told to trust that the computer knows best, and if it says raise attainment by 0.4 per cent, then raise it you must.

AI-driven mistakes?

But what if the AI has spotted patterns that don’t make sense?

We’re told that it looks at “similar schools”, but with no way of knowing how it’s reached that conclusion.

No doubt primary and secondary targets will be different, and I’m sure that basic deprivation data is a factor. But who knows what else the AI has spotted among its patterns.

The trouble with data is that there’s almost always a possible pattern.

Nearly a decade ago I highlighted the farce of comparing groups when I discovered that pupils living in even-numbered houses achieved less well than those in odd numbers.

Of course, it was a nonsense, but the gap in the statistics was greater than for boys and girls, or one of the many other groups we insisted on comparing.

Maybe the mystical AI number-cruncher spotted that, overall, schools with longer names have higher attendance data. Or perhaps those with “village” in their name are top of the table.

Neither of these conclusions are likely to help the school down the road trying to overcome entrenched school avoidance.

Who knows what advice you might get from one of the schools with great attendance? Perhaps your best bet is “hope for a better cohort next year”?

Michael Tidd is headteacher at East Preston Junior School in West Sussex

You can now get the UK’s most-trusted source of education news in a mobile app. Get Tes magazine on iOS and on Android