Another week, another article claiming that very soon teachers will be vacating their jobs to make way for the artificial intelligence (AI) of the future.

Now I’m not yet a fully paid-up Luddite - I am constantly using computers and associated technology for school business - but I can see some serious defects in the assumptions behind the latest glossy claims.

Visit any school first thing in the morning - when photocopiers are under greatest pressure - and you will notice an array of “out of order” notices, lights flashing as the machines can’t cope without more toner, etc. I often ponder how long I would be in a job if I took as long to get going on a task as the photocopiers do - and they are utterly inflexible.

School wi-fi systems are much better than they used to be. There was a time when it was impossible to show a PowerPoint or a YouTube video, let alone use all the wonderful commercial software for maths or geography. But even in this day and age, the internet goes down, power cuts happen and the plug gets pulled in all kinds of other ways. 

The technology may sound appealing, but let’s not forget that it is expensive in terms of software and hardware. Nor can it be 100 per cent reliable. Maintenance and training costs often attend the original contract. By contrast, teachers are more cost-effective and they can be directed much more flexibly.

The limits of artificial intelligence

Don’t get me wrong. There are many things I would love to turn over to computers. Trip software could be extended to be voice activated. I would find it almost congenial to be asked the relevant questions by a helpful program. It would be a breeze to go through all the risks attending anything from a walk down the road with a group to a residential on the other side of the world. Or even, open the doors for the lines and lines of children and pupils on their way to afternoon assembly - as done by the robots created by Boston Dynamics. The savings in time and frustration would be well worth the outlay.

Using AI to count the books out on loan, fill in the spreadsheets and churn out bills to replace lost copies would be welcomed with open arms. Similarly, AI to convert exam results in the summer into value-added calculations showing numbers above and below predicted grades - and other metrics - would save the departmental abacus counting these manually. It would be enough just to have exam board and school software that talks to each other.

Voice-activated software does exist, but it is in its infancy. Dragon is the only program I know of and it needs a certain amount of training and adaption time before it is fully operational. But I would be happy to have such a facility available to devise emails to parents, routine correspondence and for filling in the many forms that are now in the domain of the middle leader - since there are no signs of the hunger for paperwork diminishing in the near future.

A voice-activated register which logs itself on while the teacher gets the students down to work or holds important individual feedback dialogue would also be welcome.

Nuance needed

When I envisage the future, I remain doubtful as to whether it would ever be possible to instil the nuance needed to conduct lessons on English literature. By all means have examinations conducted via word processing - handwriting gets harder to decipher every year.  It takes considerable human ingenuity on the part of all examiners to cope with poor handwriting, such advances would make it much easier to mark online scripts more accurately and remove any bias caused by types of handwriting. But given the immense difficulty of even the most senior of examiners to agree marks for unusual scripts at GCSE and A level, we would be irresponsible to let loose any AI on the scripts of whole cohorts.

I have no objection to AI in principle. There are many chores suited to its current developmental stage that could be delegated to machines. But teachers have a unique skill-set, depth of knowledge and breadth of intelligence that is better suited to directional activity. Having machines pick up the chores would save hours on learning administrative programs - and hours spent bean counting. In this way not only would we would stop turning teachers into mechanistic administrators, we could save their energy and intelligence for the core tasks of teaching, the ones that make the most difference to students. 

Yvonne Williams is head of English in a secondary school in the south of England and has co-authored an article How accurate can A-level English literature marking be?, published in the current National Association for the Teaching of English’s English in Education journal 

Want to keep up with the latest education news and opinion? Follow Tes on Twitter and like Tes on Facebook