����Vlog

The conversation about how AI will reshape education has become inescapable - whether it’s being used for marking, lesson planning or even the Department for Education planning to roll out its own AI tutoring.

The final step on this path, some fear, is AI-powered humanoid robots taking over from teachers and teaching assistants.

But could this science-fiction future really happen?

It may actually be closer than we think. Earlier this summer, American company it would deploy a humanoid robot to schools in the Salamanca City Central school district in New York state.

The robot was set to use “natural conversation, facial expressions and real-time interaction to create engaging, hands-on learning experiences”, Realbotix said, and the aim was “to encourage student participation”.

Realbotix CEO that pupils would have a unique code to give to the robot, which would then be able to access data on their learning and previous interactions with that student.

He added that the robot would know the school district’s curriculum, and be able to help teachers out during lessons.

The scheme was part of a planned wider offering from the company to the school district, giving teachers and students access to an AI-powered teaching assistant and at-home tutor.

However, after a backlash on social media, part of which pointed out that Realbotix’s separately operated sister company, Intima LLC, has an 86 per cent ownership of RealDoll (a manufacturer of sex dolls), the .

In a statement, Realbotix said: “Realbotix LLC and Intima LLC are different legal entities that are independently operated. The two entities do not share employees, products, inventory, manufacturing resources or technology.”

AI robots in schools

Yet despite these particular plans being paused, this is not the end of the evolution of this technology. In fact, robots are already being introduced into schools in other ways.

Over in San Diego, a chain of charter schools has bought two AI-powered humanoid robots, reportedly as part of a pilot to look at how .

Meanwhile, in the UK robots have been used to support children with additional needs, says Rose Luckin, professor of learner-centred design at University College London, and these have been genuinely beneficial.

For example, the University of the West of England has used a robot called Pepper in work with autistic pupils, and the University of Hertfordshire developed a robot called Kaspar to help autistic children practise social interaction.

For several years, some schools in the UK have also been using “telepresence” robots to enable students who cannot attend school physically to feel like they are part of the class while learning remotely.

These robots, which were not designed to look like humans, can be controlled by a student at home, using an iPad. The child can make the robot look around the room while they watch a livestream of the class.

This technology has been used for children with long-term illnesses and, more recently, in Wakefield and the Wirral. The pilot in the Wirral in 2024 reported a 21 per cent improvement in attendance for pupils avoiding school.

Beyond this, physical robots haven’t yet made their way into British classrooms - but AI chatbots and AI-powered screen assistants are very much present, with Google DeepMind reportedly seeking to build a chatbot to support teachers as of late last year.

Other models that take this technology even further are also already operating in schools. For example, in Alpha Schools in the US, students spend two hours a day on academic learning through AI-based apps, and then spend the rest of the day on life skills and projects. The idea is that while the AI takes care of the traditional lesson content delivery, human “guides” - many of them former teachers - can focus on motivating students and giving them emotional support.

So, with this technology moving at pace, could we actually see humanoid robots filling teaching assistant roles any time soon?

Expecting a robot to teach ‘doesn’t make sense’

Professor Neil Selwyn, a professor in the faculty of arts at Monash University, Australia, and honorary research fellow at the University of Oxford, is sceptical. He calls humanoid robots a “gimmick”.

“In reality, there are many reasons that humanoid robots simply don’t work in classrooms - from the obvious health and safety issues, through to the fact that these machines simply don’t work very well,” he says.

“Any school using these sorts of robots will need real-life teachers in the room doing a lot of work to make sure nothing goes wrong,” Selwyn continues.

“A teacher will need to be on hand to clarify and reinterpret what the machine is saying and doing, to deal with any unexpected detours and interruptions. No lesson runs like clockwork and humans are a lot better than AI at thinking on their feet, improvising and dealing with emergencies. Plus, real-life teachers will know what the students in the room actually need at that time to learn effectively.”

He adds that teaching “is not a predictable, scriptable, robotic process” and so “expecting a robot to be able to teach does not make sense”.

Ultimately, Selwyn warns, “rather than making teachers’ work easier, technology like this will require teachers to do a lot more work around the edges in order to keep the show on the road”.

On top of this, there is a more general concern about creating a “two-tier system”, where schools with more resources can do the additional work required to make sure that new technology is used effectively, to give teachers more time with children, while schools under financial strain use the same tools “to teach on the cheap”, Luckin says.

Realbotix’s robot would have been deployed to a community where 79 per cent of students are economically disadvantaged, she points out.

The DfE’s own AI tutoring programme will be targeted at disadvantaged students, and a government spokesperson has explicitly said that it is aiming to give disadvantaged kids an AI-version of the support their wealthier peers can get from private tutors.

“New technology is often trialled first on the pupils with the least power to refuse it, and the promise of access can slide into a substitute for the human teaching that disadvantaged children need most,” Luckin says.

AI can be helpful if it allows a working teaching assistant to extend their reach, she adds.

But if new technology is used to justify larger classes or fewer staff, then ultimately “technology will widen the gaps it was meant to close”.

��