As Felix Simon recently argued, claims that ‘AI will replace profession X’ tend to come from people with very little understanding (let alone firsthand experience) of working in these professions. This is certainly the case with hype around AI replacing accountants, doctors and judges. The same goes for claims around receptionists, administrative assistants and other lower-status jobs soon being rendered obsolete by AI and other forms of automation.

In truth, these are all jobs that demand endless amounts of relational and emotional work, complex tacit knowledge, high-level improvisation skills, and many other nuances that transcend the formulaic ‘tasks’ that workplace AI is being developed to take on. As such, we need to remain highly sceptical of ‘AI replacement’ narratives being pushed by an IT industry desperate to find convincing use cases for its products.

The myth of AI replacement in education

This need for caution certainly applies to current debates about teachers being replaced by AI. Indeed, this is an argument that nearly every high school teacher in our DP25 study has found an utterly preposterous argument to make.

Nevertheless, there are plenty of others outside of the teaching profession willing to promote ideas of ‘AI tutors’, robot teachers and personalised learning systems taking care of students’ learning. As Sam Altman (2024) puts it: “our children will have virtual tutors who can provide personalized instruction in any subject, in any language, and at whatever pace they need.”

Of course, pronouncements from the likes of Sam Altman might well be dismissed as self-interested efforts to sustain the AI investment bubble. Yet, the idea of AI replacement lurks around the edges of many more sober conversations around educational AI.

Usually, AI proponents are careful to not explicitly proclaim the complete disappearance of teachers – after all, these are technologies that need to be sold (at least in the short term) to schools and universities. Yet beneath current messaging about AI being used to ‘help’, ‘support’ and ‘free up’ teachers from onerous aspects of their work are implicit beliefs that AI can take on the main tasks traditionally undertaken by expert professional educators.

For people working outside of education, such thinking might well seem to make good sense. If AI can take care of lesson planning, content presentation, student assessment and feedback, then most students will only sporadically require support from a human (most likely in the guise of classroom assistant or critical friend rather than expert teacher). 

This logic is pushed even harder when it comes to older students – particularly those in tertiary education who are assumed to be capable of self-regulated and self-directed study.  Here, then, the need for highly paid, expensively trained expert ‘lecturers’ seems even more obviously under threat.

The difficulty of talking about AI and teaching when everyone considers themselves to be an expert on teaching

In many cases, such enthusiasms are motivated by the genuine belief that AI can perform teaching tasks at far greater speed, scale, efficiency and insightfulness than is possible with human teachers – particularly in terms of individually attending to students’ specific needs and supporting ‘personalised’ forms of learning. 

In part, this idea of AI doing a better job that a human teacher gains credence because nearly everyone who went to school as a child tends to consider themselves to have ample experience of what teachers do. Often this is compounded by parents and grandparents developing opinions based on their own (grand)children’s purported experiences.

All these experiences are highly partial, often wildly out-dated and impressionistic. Nevertheless, teaching is unique as a profession that most people consider themselves eminently qualified to have strong opinions about and/or consider themselves to be capable of doing if so inclined (quite unlike the job of being a doctor, judge or accountant). As Ira Socal and colleagues (2009) put it:

“Everyone is an expert on education and its particular, dominant subset – school. Everyone who has either attended school, taught at school, had their kids at school, managed school, funded school, even avoided school knows what school does. Unlike any other public institution, we can quickly produce an opinion on what schools should and shouldn’t do. Scores of politicians, business leaders or (other) powerful pundits who arrive on the scene claim the credential of knowing how to run schools. Many of these self-proclaimed experts are widely interviewed and financially supported”.

This lack of respect for teachers’ work is compounded amongst those in the IT industry who consider themselves to have developed their tech know-how despite (rather than because) of their schoolteachers. The tech industry abounds with origin stories of talented young people held back by formal education and only flourishing through self-regulated forms of extra-curricular learning. Indeed, the self-taught and self-made software developer is a recurring part of the mythologising of Silicon Valley luminaries from Steve Wozniak to Jack Dorsey.

All told, the idea that AI might replace teachers (and the accompanying implication that this might well be a good thing) is an easy message to sell to an IT industry and general public with relatively low opinions of the teaching profession.  As such, it is also an appealing idea to politicians and policymakers desperate to be seen to be doing something to address the intractable challenge of improving public education.

So, what needs to be done?

The problem, of course, is that teaching is a highly complex job and that the structural shortcomings of public education systems are not easily fixed through the development of an AI tutor or learning system. The imposition of AI into schools is largely forcing teachers to engage in additional and/or worse forms of work. Efforts to support/augment/replace teachers are certainly not seamlessly leading to better forms of schooling and education.

All the hype around ‘AI will replace teachers’ is as unfounded as with any other profession, and the teachers in our research project are well justified in being unswayed – if not bemused – by such statements. Yet, this is not to say that the education community can simply ignore the current ‘AI will replace teachers’ hype. These are not narratives that look like naturally dying down, and it is important for the education profession to stand up and actively make the strongest case possible for why teachers cannot (and should not) be replaced by AI technology.

As Adam Greenfield (2017) reminds us, it often does not matter whether hype around digital technologies is correct – what matters is if someone believes it to be true and then acts upon this belief. Clearly, there are many people in industry and policy circles who are hostile to the idea of expensively trained expert professional educators who have jobs for life, pension rights and union protection. Clearly, many of these people are currently happy to push ideas of AI replacement as a way of undermining the status of the professional teacher. 

These narratives around AI replacement in the classroom now need to be loudly challenged and systematically debunked … with teachers being the ideal people to begin leading counter-conversations around how the work teaching and learning is far more complex and complicated than non-experts might believe. The time for speaking back to the AI hype in education starts now!

REFERENCES

Greenfield, A. (2017). Radical technologies. Verso 

Socal, I,  Lazic, T. and Thompson, G. (2009). Why is everyone an expert on education?