Two conferences this week and I got the now obligatory conference-chat about robots at both.

Andreas Schleicer, the head of the PISA tests and a perma-attendee at conferences, showed a slide at the Whole Education Conference stating that “the kind of things that are easy to teach are now easy to automate, digitise or outsource”.

The room nodded, enthralled. But is it true?

Maths seems easy to automate. Calculators are a simple form of computer. Yet sums are not easy to teach. Pupils spend around an hour a day doing maths for 12 years, but most would struggle to quickly divide 289 by 17, whereas it took google a fraction of a second to tell me the answer (it’s 17).

Thinking about it more I couldn’t help wondering what things really are easy to teach? Every child soon learns how to turn a handle, as my friends with toddlers are horrified to find. Hence we might imagine this is an easy thing to teach.

Robots, however, struggle with handles. Boston Dynamics recently unveiled a video of a “robot-dog” which took eight years to programme to – gasp – OPEN DOORS. Plus the rigmarole that poor botdog must go through to make its escape suggests that we are very far from translating things that humans find easy into things that robots can do.

All of which may explain the negative reaction at the second conference of my week. At the Head Teachers’ Roundtable Summit, Robert Halfon, the chair of the parliamentary education committee, said he was concerned that 30 per cent of jobs would soon be done by robots. You could almost hear a hiss in the room as the words left his mouth.

The line “we could retrain workers who lose their jobs” is not very natty as a conference slide or a soundbite

Yet in fairness to Halfon, there is evidence in his corner. Bernard Marr, an author and public speaker, who sells a lot of services prepping people for a digital future, recently released research showing that 35 per cent of jobs are indeed at risk of automation.

But less-invested experts put the number lower. US market researchers Forrester guess six per cent, while McKinsey reckons that 60 per cent of jobs could see a third of their activities automated (which seems a roundabout way of saying that just 20 per cent of jobs are at threat).

But then McKinsey’s report also says the number of people needing to switch jobs will be between 75 million and 375 million. That’s equivalent to everyone in the UK and Ireland losing their job at the lower end, and the entire US population losing theirs at the top end. It’s so wide a prediction that it’s basically useless.

Ultimately we don’t know how the future will go. Daisy Christodolou, who wrote The Seven Myths of Education, tells the story of how her grandparents’ handkerchief market stall, which they ran for 50 years, was wiped out in the 1990s by Sunday trading laws and Kleenex. No amount of worrying about the future when her grandparents were at school would have solved that problem.

Instead of gnashing our teeth and trying divine future worlds, I was pleased to hear Emran Mian, a senior civil servant at the Department for Education, talk about the National Retraining Scheme – one of the government’s manifesto pledges last year. As Mian explained it, the scheme will look at areas where jobs are at risk of automation and create training packages to reskill workers. Job done, and not a single school policy needs to change.

Unfortunately, the line “we could retrain workers who lose their jobs” is not very natty as a conference slide or a soundbite. So here’s my suggestion to everyone at the DfE: rebadge the National Retraining Scheme as the Anti-Robot Defence Plan.

Everyone loves robots. It’ll be the talk of the conferences in no time.

Laura McInerney is Contributing Editor of Schools Week