Unlock the Editor’s Digest for free
Roula Khalaf, Editor of the FT, selects her favourite stories in this weekly newsletter.
As students return for a new school year, artificial intelligence is beginning to seep into the classroom. In at least one London private school, a pilot group will begin studying core subjects for GCSE exams in “teacherless” classrooms using AI-powered systems, supervised by “learning coaches”. The disruption of teaching by technology is, though, not universally welcomed. In South Korea, a plan to introduce AI-powered digital textbooks in schools in 2025 has sparked a backlash among wary academics and parents.
As in other fields, AI is in reality unlikely to replace teachers any time soon, and nor should it. Research suggests the best learning is social, involving interaction between teachers and students, and between students themselves. This is one reason why “massive open online courses” — open-access, web-based courses aimed at wide participation — have not lived up to the hype around their arrival in the 2010s.
AI is also unlikely ever to be able to substitute for the judgment, motivation, guidance and pastoral care a teacher can provide. But since students will enter AI-equipped workplaces, schools must prepare them to navigate that world, including through how they learn. And AI holds the prospect of empowering teachers — giving them more time to do what they do best.
The technology can, for example, help with the hours of grunt work, sometimes unpaid, that teachers do outside the classroom. Educational technology companies are already launching products using generative AI to help teachers prepare lesson plans and presentations, and tasks for students, though, like AI elsewhere, these must be policed for “hallucinations” and bias. Increasingly sophisticated systems are being launched, too, that can mark tests and homework, even providing feedback on written work.
Ed techs are also developing AI-powered tutors that can open the way to giving students more individual attention by tracking their progress and understanding and providing tailored support. South Korea’s digital tablets, for example, are billed as being customisable so that both fast and slow learners can be assessed by software and given specific, AI-generated tasks. Teachers must, though, oversee students’ progress and complement online learning sessions with plenty of class discussion and joint activities.
A headache for teachers has been ensuring students do not outsource learning completely to AI, and use it to write their essays and maths answers. Some schools are expanding the concept of “flipped learning” and moving away from asking students to produce written work at home where AI is easy to deploy. Homework is instead used to study learning materials — potentially AI-aided — and classroom time to deepen and test students’ understanding through discussion, problem-solving and some supervised written work.
AI-assisted tools can make learning more efficient. They may also enable subjects to be taught in more creative ways. Students can have conversations with avatars in foreign languages, visit the International Space Station using virtuality-reality headsets, or use AI software to co-create music and art.
Indeed, the advent of AI not just in education but across the workplace will necessitate a rethink of exactly what schools teach and how they teach it. The instant access to information technology offers ought to facilitate a move away from rote memorisation and towards developing students’ ability to apply their knowledge and skills.
By making teachers more productive, AI-assisted education offers the possibility of compensating for shortages of skilled staff, especially in poorer regions and countries. But in order to reap the full benefits of the technology — without undermining existing educational standards — teachers, schools and governments will need to adapt too.
