Education in the Age of AI: Ideas from business
Introduction
In my work I spend a lot of time around software, technology, and now increasingly AI. I see first-hand how quickly it’s changing the way businesses operate, and I often get asked about what that means for the future. Education always comes up in that conversation — and as a parent with school-age children, it comes up at home just as often.
This article isn’t written as an academic or as a teacher — that’s not my background. It’s written as someone who uses AI every day in business and has started to think about what that means for schools. Teachers invest years of effort into their lesson plans and materials over their careers, and with the changes that AI will bring, the work and experiences of teachers can be utilised in new ways. The question is how AI can build on that work and enhance it.
What I’ve tried to do here is organise some ideas and experiences from business into a narrative - as a thought experiment. The first section sets out a broad view of how teaching has evolved in ways that lend themselves to utilising AI, detailing the challenges AI presents, and what opportunities it opens up. The second section then goes into practical detail about how the things we use in business today could work in classrooms too. Perhaps this article sparks some debate — about how we can use AI in ways that keep teachers at the centre, prepare students for a world where they may never be the smartest in the room, and bring forward the human skills that matter most.
Part One: The Changing Role of Education in the Age of AI
The existential shift
For most of history, the classroom has worked on the simple assumption that the teacher is the one who knows the most. That assumption gave teachers their authority and their central role. But with the rise of generative AI — systems like ChatGPT and NotebookLM — knowledge, facts, and connections are now accessible in seconds. In many domains, AI already outpaces humans in speed, and recall. The teacher may no longer be the one with the deepest store of information.
This alters something important: what is the role of a teacher, when the teacher is no longer the most knowledgeable? For students too, the same challenge looms, students will rarely be the smartest entity in any conversation. They will enter a world where intelligence is distributed, where AI is always in the room, and they will need to learn how to bring the distinctively human skills that make them valuable. The challenge to teachers applies equally to their students.
From repetition to co-creation
Education has moved through major shifts before.
Behaviourism treated learning as the transmission of knowledge from teacher to pupil, reinforced through repetition and reward. It was effective for memorisation but too limited to foster deep understanding or creativity.
Constructivism reframed learning as something learners actively build for themselves. Knowledge is constructed, not passively absorbed. But this could sometimes feel mechanistic or disconnected from social interaction.
Social constructivism places emphasis on dialogue, collaboration, and the co-creation of meaningful outcomes. Here, the teacher facilitates live interaction, supporting not only factual learning but also skills such as research, critical thinking, problem solving, and creativity.
Each teaching evolution widened the scope of what education could achieve. Social constructivism is closest to what many classrooms aspire to today: students working together on shared problems, guided by the teacher, building not just knowledge but human skills along the way.
Yet from what I have learned, even this approach has its challenges, particularly around measurement. Dialogue and collaboration are powerful but difficult to capture in ways that schools and assessment systems recognise. Because of that difficulty, schools often default back to tasks that are easy to measure — written work, copying from the board, long essays, filling PowerPoints. The result is that students spend five minutes on research and 25 minutes writing it up, when the real value must lie the other way round.
This tendency undermines the very skills that social constructivism was meant to foster, and it distorts creative subjects most of all. Art students are asked to write about artists instead of creating art. Music students might learn the facets of instruments instead of making music. These compromises show how the pressures of measurement can flatten education into written propositions, even in domains where expression, creativity, and participation should be at the centre.
Why human skills matter in the age of AI
If AI is now “smarter” in the sense of holding more facts and generating faster outputs, what remains uniquely human? In my work I find that the answer lies in the skills AI cannot easily replicate: generative insight, imaginative leaps, good judgment, originality, curiosity, and the ability to connect across domains with purpose and empathy.
From my wider reading I’ve understood that international bodies such as the OECD and the World Economic Forum have consistently identified the most vital “21st-century skills” as collaboration, creativity, problem solving, critical thinking, and adaptability. Research skills matter just as much — not simply gathering information, but questioning its relevance and quality. Curiosity drives inquiry, creativity enables new connections, problem-solving applies judgment, and collaboration ensures knowledge is social, not solitary.
Together, these skills underpin what cognitive scientists are calling relevance realisation — the ability to discern what information matters, when, and how to apply it to create value. These are the skills the world will need in abundance when humans are no longer the smartest in the room.
In business, this shift is already visible, as a leader I try to role-model curiosity and creativity, encouraging better questions, exploration of alternatives or ‘thought experiments', and frame problems in new ways. They know AI will surface the data and generate the summaries; what matters is the human capacity to steer, interpret, and decide. Schools now face the same opportunity: to put these human traits at the centre of how young people learn.
The AI inflection point
This is where AI becomes more than a challenging subject — it becomes an enabler. In business, transcription and summarisation tools already run in the background, capturing meetings, surfacing themes, and creating a record without human effort. Leaders are beginning to show that by using AI to strip away repetitive work and administrative burden, they can free more time for curiosity, creativity, and connection. These behaviours are increasingly the ones that drive value.
In the classroom, the same logic applies. AI can capture dialogue, highlight key points, and generate resources for reflection. It can act as a prompting partner, helping students explore questions, spot gaps, and expand ideas. It can generate multiple formats — from mind maps to quizzes to podcasts — giving students different ways to engage with content. It can even support assessment by clustering responses, suggesting feedback, and reducing administrative burden, while still leaving judgment to the teacher.
The key is to embrace AI not as an intruder, but as a co-worker. The teacher’s role shifts from being the sole source of knowledge to being the most creative facilitator in the room: the one who designs the questions, curates the dialogue, and models how to use AI critically and responsibly.
We can argue all day over whether or not AI is a good thing, but the fact is it is here and I believe we are better off embracing it’s exciting potential.
Part Two: Making AI work in the Classroom
The following section draws on experiences from business, where tools like ChatGPT and NotebookLM are already maturing and proving their value. The exact platforms will change over time, and schools will eventually make their own choices about which to adopt. What follows is not the only way forward, but one possible recipe for how teachers and students could begin experimenting with these tools in practice.
Preparing for lessons - a teacher workflow
Teachers have already invested years in building their materials, and that work remains invaluable. A first step might be to preserve and build on it. Existing resources — from worksheets to handwritten notes — can be digitised using nothing more than a smartphone camera.
Once digitised, the material can be uploaded into ChatGPT as an information augmentation tool. This is not about “getting more text back,” but about using the AI to help organise, check, and strengthen what already exists. For example, a teacher could upload their notes on the Battle of Hastings and ask ChatGPT to:
organise the events into a clear timeline,
group information into themes such as strategy, leadership, and context,
flag likely misconceptions that students may hold, and
expand shorthand notes into coherent teaching points.
This turns existing material into a structured and relevant resource. The teacher can then refine it — deciding what to keep, what to cut, and how to frame it for their class. This is an act of relevance realisation: drawing out what really matters for their students at that moment.
From there, the refined and enhanced versions of lesson notes can be uploaded into NotebookLM, which has provided a rich frame for learning in business already. With curated content as a starting point, the system can then generate targeted outputs — mind maps, quizzes, podcasts, explainer videos, or interactive tasks — aligned to the teacher’s intentions.
Attending a lesson - a student workflow in class
Students can follow a similar workflow, though their starting point is different. Instead of long-term teaching materials, they begin with their own work: class notes, handouts, assigned readings, or sources they have researched themselves.
These materials can be uploaded into ChatGPT. As with the teacher’s process, the aim is to structure and enrich what they already have. The AI might turn their bullet points into a coherent summary, show how ideas connect in a hierarchy, or surface gaps where further evidence is needed.
Students then refine the AI’s draft, deciding what is useful, what is missing, and how to frame it in their own words. In doing so, they practise relevance realisation: the skill of judgment and prioritisation. Once refined, their material can move into NotebookLM, where outputs like quizzes, podcasts, or mind maps provide different ways of consolidating and demonstrating understanding.
Co-creating in the room: facilitation, modelling, and on-screen practice
Another important part of this approach is for teachers to use their refined content and student contributions together in real time. They might project a ChatGPT thread containing their structured notes, set out the learning goal, and show the class how they prepared it. This frames the tool as something transparent and shared.
Teachers can then model these new ways of working by prompting out loud: why they are giving context, what format they are asking for, and how they will use the result. A simple frame such as why / what / how can help students see the process. Students might then suggest prompts, critique outputs, and even take turns as a “prompt lead,” coming to the front to write the request and explain the intention behind it.
During this process, transcription tools such as Otter.ai could run quietly in the background, with the class informed in advance. We do this at work. The transcript serves as a record and reference: capturing turning points, key questions, and unresolved points. Afterwards, it can be summarised, used to seed next-lesson slides, or form the basis of a short quiz that reflects the discussion.
To close the co-creation phase, teachers might ask the model to summarise the class dialogue under the key headings used in the lesson, using excerpts from the transcript to anchor it. These materials, combined with the refined preparation, can then be moved into NotebookLM and be used within videos, podcasts, mind maps, quizzes as learning materials. The result is a package that contains the structured content, the class discussion, and the distilled notes — from which more outputs can be generated as needed.
Homework and independent study
Homework can be reframed using the same workflow. Rather than spending most of the time writing, students could spend more time on research and refinement. For instance: “Upload your Hastings notes into ChatGPT, structure them into three themes, refine the results, then generate a NotebookLM quiz and test yourself. Submit both the quiz and a short reflection on what you learned.”
This approach flips the time balance: less time on repetitive writing, more on thinking, refining, and creating.
Embedding and follow-up
In a subsequent lesson, the teacher could revisit both their own enriched notes and the student-refined versions. Together, the class can reflect: Which details mattered most? What did AI highlight that we overlooked? What did we cut that it over-emphasised?
This cycle builds a culture of co-learning, where teacher and students follow the same discipline: start with what you have, use AI to structure and expand, apply human judgment to refine, and then create targeted outputs for deeper learning.
Transcription as measurement
Having read that the challenge of social constructivist learning has always been assessment caused by dialogue and collaboration being hard to capture. Transcription tools now provide one possible solution.
By running in the background, they create a record of who said what, what ideas emerged, and how thinking developed. Teachers do not need to read every word, but transcripts can be summarised automatically, searched for key points, and revisited when needed. Over time, they build a record of engagement — not just of memorised facts, but of reasoning and collaboration.
This mirrors business practice, where transcription has become routine in meetings. There, transcripts are used to generate summaries, action points, and records of key insights. In schools, when paired with NotebookLM outputs, they could provide a workable way to make social constructivist learning more measurable — capturing interaction without reducing it to rote writing.
Assessment and feedback
Traditional assessment rewards memorisation more than creativity or collaboration. By using AI tools, schools could begin to broaden this picture. ChatGPT might analyse drafts, highlighting strengths and raising questions. NotebookLM could generate fun quizzes for developing mastery of a topic. Transcripts could provide evidence of reasoning and participation.
Together, these elements create a more balanced view of learning. Teachers remain the final arbiters, but the burden of capture and organisation is eased.
Teacher role reimagined
For teachers, this model keeps their years of preparation central while offering ways to extend it. Their role shifts from being the only source of knowledge to being the most creative facilitator: modelling how to refine content, question AI critically, and connect material to meaning.
For students, the process mirrors this. They learn to work with AI as a partner, not a shortcut, and to bring forward curiosity, critique, and creativity. This is particularly important in creative subjects, where AI can free teachers from the pressure to turn every activity into written outputs. Instead of writing about artists, students can create art. Instead of cataloguing instruments, they can make music.
School-level implementation
Practical adoption will depend on infrastructure and culture. Schools need access to the right tools, devices for students, and agreed policies on ethical AI use. Teachers may need training in prompting, facilitation, and lesson design.
A phased approach may work best: start small, experiment in one subject or year group, and scale from there. Equity of access is essential if benefits are to be shared fairly.
Example workflow: Battle of Hastings
The teacher scans their notes and uploads them to ChatGPT, asking it to structure them into timelines, themes, and expanded explanations.
The teacher reviews and refines this version, making decisions about relevance and age-appropriateness.
The refined material is uploaded into NotebookLM to generate outputs such as a mind map, quiz, or podcast script.
In class, the teacher begins with the question: “Why did William win?” and models prompting, while students upload their own notes — including any additional sources they’ve researched — into ChatGPT.
Students refine their structured notes and move them into NotebookLM to create outputs or reviews the content produced by the teacher.
Transcription runs in the background, creating a record of the class discussion.
For homework, students repeat the workflow: structure and refine their notes, generate a NotebookLM output, and reflect on what they learned.
In the next lesson, teacher and students revisit both the transcript and outputs, embedding the learning and reflecting on the process.
Conclusion
Education has always evolved, each time expanding the scope of what learning could be. AI now presents new opportunities. It challenges the assumption that the teacher is the smartest in the room, but it also provides tools to deepen enquiry, capture dialogue, and support assessment.
The opportunity is not to replace teachers or discard the materials they have built, but to amplify their work and free them to focus on the parts of teaching that are most human. Students, meanwhile, learn the skills that will matter most: curiosity, critical thinking, creativity, collaboration, and the ability to evaluate and apply AI outputs responsibly.
If schools, like some workplaces embrace these possibilities, they could quickly become more adaptive, modern, and flexible — preparing students not only for exams, but for a world in which AI is always present, and where the most valuable human contribution will be to bring meaning, creativity, and connection.