The past 20 years have seen the greatest technological revolution in history. Almost every part of society has been transformed. Hospitals can now print 3D models of patients’ brains; militaries can blow up targets with pilotless drones; and my kids can use voice activation to play any song ever recorded.

Yet schools have been left remarkably untouched, despite regular prophecies over the past few decades that radical upheaval was just around the corner. Of course, technology has had some impact; many schools now have virtual learning portals; shared resources have made lesson planning easier; and some of the most interesting developments in assessment - such as comparative judgement - require technology to work.

But on the whole, lessons look pretty much the way they did 20 years ago. This is because no technology has come close to replacing the relationship between the pupil and teacher. Sure, teachers now dress up the transmission of information with some digital window-dressing, but using Minecraft to explain fractions is no different to using a textbook.

To my untrained eye, a lesson involving tablets looks a lot like one without tablets apart from the 10 minutes wasted faffing around with the tablets.

Self-direction or misdirection?

Moreover, the suggestion that the internet would allow students to self-direct their learning without a teacher has proved misguided. The problem of novices self-directing is that they don’t know what to direct themselves at. Unless they have high motivation, why are they going to do anything at all?

So teachers have proven their worth. Given that, are we likely to see anything change? At the margins, I think we’ll see cash-strapped schools using technology to cut costs. There are, for instance, already a few “Uber for supply teachers” companies looking to undercut agencies. And surely we’ll move to online examinations, given that the current administration-heavy process costs schools £300 million a year.

No tech has come close to replacing the pupil-teacher relationship

But the really big question is whether anything can challenge the primacy of the interaction between a human teacher and pupils. The obvious answer seems to be no. A robot can spray paint a car but it can’t begin to replicate the hundreds of complex interactions teachers have every day. In the past few years, though, we’ve seen huge developments in “deep learning”, with computers using general learning algorithms to work out how to beat humans at everything from Pong to poker without any specific programming.

Some argue that this will lead to computers that can do pretty much any human job within the next few decades, raising the spectre of robot teachers and unemployment. But it seems more likely that computers will, instead, continue to get better at specialist tasks, without developing a general emotional intelligence. In practice, augmented intelligence, where humans use computers with these skills to improve their own abilities, seems to hold more potential. Scientists have trained computers to spot patterns in how pupils learn in online tuitions. By identifying when students are confused, they can give prompts to teachers.

If this sounds like hell, I wouldn’t worry too much. It’s taken most schools more than a decade to work out how to make use of electronic whiteboards, so augmented teaching is a little way off. But if it could help teachers improve their practice, we shouldn’t be afraid.

Sam Freedman is executive director of programmes at Teach First and a former government policy adviser