AI in the Classroom – The Upending of Traditional Education

Photo by Element5 Digital on Unsplash

On a recent episode of the New York Times' Hard Fork podcast, I was startled to hear about an "AI-first" school that has used AI to reduce academic teaching time to a mere two hours per day and has redeployed teachers to be mentors and guides, rather than deliverers of content, policing, and grading. What happens after the two-hour academic class? Lots of fun, apparently – collaboration games, motivation exercises, and life skills activities. The school boasted that students are scoring in the 99th percentile nationally for its core curriculum. Oh, and no homework. Ever.

We'll get back to the veracity of these claims in a moment. Most of the discussion about AI and students has been around the use of GenAI to circumvent the hard work of doing homework and essays and reading, by passing that job over to a chatbot, which generally does fine work on their behalf – at least if you are a student looking to shirk responsibilities and don't object to a bit of plagiarism. Schools and teachers, unsurprisingly, were rather irritated by this and instituted various fightback campaigns ranging from threats to AI bans to punishments to technology-based chatbot plagiarism detectors (the latter now generally agreed to be hopelessly inadequate).

Some educators are now realising that this is all a waste of time. Students are going to use GenAI wherever they can: at school, at home, in the bathroom. One current study has 63% of US high-school students using AI in ways that they know contradict school policy. There is no way to stop it, and it is becoming apparent that perhaps it should be encouraged within carefully designed guardrails.

And so a new crop of education initiatives is blooming where the AI designs the course plan, does the teaching, the monitoring, the personalised assistance, the remedial support (where appropriate), as well as the evaluations and grading. This leaves the teacher free to concentrate on softer skills like coaching and mentorship, making the school a more joyous and expansive experience for both students and teachers.

How is it going? The school I mentioned earlier is Alpha Schools, which now operates three campuses in the US. It is one of the better-known AI-first ventures. Its model is mixed-age microschools using adaptive platforms so students "learn 2x in 2 hours," then spend afternoons on projects, mentorship, and life skills. The idea is deliberately disruptive – take the tedium out of repetitive content acquisition and give time back for exploration. Independent observers and critics are understandably sceptical of headline ratios such as "2 times faster" learning; some commentators have flagged that such claims need careful scrutiny and peer-reviewed evidence.

In the podcast co-founder MacKenzie Price enthusiastically proclaims:

"We practise what's called the Pomodoro Technique. So, kids are basically doing like 25 minutes of focused attention in the core academics of maths, reading, language, and science. They get breaks in between, and then by lunchtime, academics are done for the day and it's time to do other things.

"So in the afternoon is where it gets really exciting because when kids don't have to sit at a classroom desk all day long, just grinding through academics, we instead use that time for project-based, collaborative life-skills projects. These are workshops that are led by our teachers – we call them guides – and they're learning skills like entrepreneurship and financial literacy and leadership and teamwork and communication and socialisation skills."

There has been some grumbling about her claims. They are self-reported numbers, and a couple of parents have been critical. Also, the school is not cheap, so the results would certainly show some selection bias. On the other hand, they do not reject children with poor academic track records, which is more than can be said for other private schools.

Still, places like Alpha have pushed the conversation from "can we use AI in classrooms?" to "how radically should we reorganise the school day around it?" – and that question is exactly the kind of policy and practice debate we need.

There are other schools too, like the private David Game College in London, currently piloting an AI-first teaching curriculum for 16-year-olds, which looks very similar to Alpha Schools, with similar sterling results. But there are only 16 students in the pilot, so the jury remains out.

What about the public-school system? Is AI just for the rich? Turns out, no. There is an impressive case study in Putnam County Public Schools in Tennessee. Facing a severe shortage of computer science teachers, the school district partnered with an AI platform called Kira Learning. What happened next was a true testament to the power of a good hybrid model: 1,200 students enrolled, and every single one of them passed the course – which would be remarkable in any subject, let alone computer science. This wasn't "AI-first" in the Alpha School sense. It was "AI-for-all," a tool that levelled the playing field and brought a vital skill to students who might otherwise have missed out. The teachers weren't replaced; they became "facilitators," freeing them from grading and lesson planning to focus on student engagement.

It is not only students that benefit. Consider AcademIQ in India, an AI-powered educational application designed to revolutionise the learning experience for teachers, students, and parents, particularly in multilingual and low-resource classrooms. Launched in 2023, the platform is India's first to be built around the National Education Policy (NEP) 2020.

The platform's primary appeal lies in its simplicity. It offers over 150 AI tools, most notably an "instant NEP-aligned lesson plan generator" that requires no prompts or technical expertise. Teachers simply select a class, subject, and topic to receive a complete lesson plan in seconds. This is a direct response to a major pain point for educators; testimonials indicate the platform can reduce weekly lesson-planning time by several hours, freeing up teachers to focus on student engagement.

What then of pedagogy – the art and science of teaching? Universities, which have a long heritage in teacher training, are notoriously slow to change curricula. If it is true that everything except the interpersonal roles of mentorship, coaching, and guidance will be outsourced to much more efficient AI teachers, what then should be the requirements of a teaching degree or certificate? It is not clear that anyone has wrestled with this question sufficiently, particularly given AI's dizzying rate of improvement.

For those teachers terrified of these coming changes, but resigned to the imminent primacy of AI tutors, there is this comfort – no more lesson plans, no more grading, no more one-size-fits-all. Just the nobler task of shaping the critical human traits of motivation, confidence, enthusiasm, collaboration, and childhood curiosity.

That sounds much more satisfying than the old way.

Steven Boykey Sidley is a professor of practice at JBS, University of Johannesburg and a partner at Bridge Capital and a columnist-at-large at Daily Maverick, Daily Friend and Currency News. His new book "It's Mine: How the Crypto Industry is Redefining Ownership" is published by Maverick451 in SA and Legend Times Group in UK/EU, available now.