Looking back, I can’t honestly say that if I’d had the chance, I wouldn’t have used AI to do my homework.
The school I went to kept such tight controls that not doing homework was never really an option. But it was always the thing I left until last. Sometimes I’d scribble it on the bus on the way in, just enough to get by.
Homework was a nineteenth-century invention, born out of necessity. Classrooms were overcrowded, literacy was the priority, and teachers needed a way to extend learning beyond school hours. Copying passages, memorising tables, repeating sums. It was efficient, measurable, and it served its purpose.
That context matters. For many students today, AI has become the fastest way to escape another dull assignment. Essays appear in seconds. Problems solve themselves. Homework collapses into copy and paste.
The Brookings Institution has just published an important study (January 2026) finding that students across multiple countries are already using generative AI to complete homework tasks, particularly those they see as difficult or low value. Many describe this as a choice rather than a necessity. They believe they could do the work, but don’t see why they should.
Brookings also highlights the risk. When AI replaces thinking rather than supporting it, students disengage from productive struggle. Cognitive effort is outsourced, and a valuable opportunity to learn is squandered.
Motivation Matters
Psychologist Mihaly Csikszentmihalyi, "the father of flow" argued that motivation appears when challenge is balanced. Push too hard and people shut down. Make it too easy and they drift away. Engagement lives in the space between the two. A kind of Goldilocks zone that encourages students to step up rather than check out.
One of the remarkable things about interactive ai learning tools such as role play is that the system can adapt the difficulty level to the individual level. It levels the playing field, no one is left behind, everyone gets just the right challenge to get into a state of flow.
That matters because AI has changed the conditions. When students use AI to complete homework for them, the challenge collapses. And when challenge collapses, motivation has nowhere to land.
The technology itself isn’t the problem. It’s how it’s used. The Brookings report raises legitimate concerns, but it also points to a way forward. With clear frameworks for how AI is used in learning, it’s still possible to protect what matters while preparing students for the world they’re growing up into.
Meeting the Demands of a Changing World
Even when I was at school, a lot of the curriculum felt irrelevant. It was hard to stay interested in maths, Shakespeare, and geography when they felt so far removed from the world I could see around me in the UK in the 1990s.
Education has always been about preparing children for the world they will grow up into. From tribal children learning by following adults, to the squires and apprentices of the Middle Ages, learning was rooted in real life.
Today, the pace of change in the world is beginning to outstrip our ability to adapt to it.
Many parents find it hard to keep up with the latest K-pop trend or the TV shows their children are watching. At a governmental level, policy often lags behind the emergence of new problems. Institutionally, too, processes put in place long ago become embedded as best practice, even when the world they were designed for no longer exists.
Bottom-Up vs Top-Down Change
Before taking office, Keir Starmer argued plainly that “the ability to speak well and express yourself should be something that every child is entitled to and should master… Oracy is a skill that can and must be taught.” Two years on, a coherent nationwide policy has yet to emerge.
But at an individual level, something more hopeful is already happening. Visionary teachers are experimenting. Just last week, I saw a LinkedIn post from an oracy teacher who had programmed Stalin into a character-based AI and put students into a live question-and-answer session. It got students talking. It turned history into a conversation rather than a worksheet.
That, for me, is the real signal. Change rarely arrives fully formed from the top down. It starts in classrooms, with teachers willing to rethink practice and use new tools to make learning feel alive again.
AI Literacy vs AI Fluency
In conversation earlier today, a colleague tried to prove me wrong by saying, half-jokingly, “Well, my Gemini says otherwise.”
I laughed out loud and replied, “That’s not what my ChatGPT says.”
It captured the zeitgeist. But I also knew something he didn’t. I’d checked the answer for myself outside the app. Fact-checking, weighing things up against your own experience, and applying critical thinking are skills that matter now more than ever.
AI literacy is knowing how to use the tools. How to prompt them, how to get an answer, how to move quickly.
AI fluency is something else. It’s knowing when to trust an answer, when to question it, and when to ignore it altogether. It’s the ability to think independently with AI in the room, not surrender thinking to it.
Allowing students to work with AI under teacher supervision, in guided practice sessions, can help train these skills. Exercises such as role-play, goal-play, critique, and co-creation, as proposed by Ethan Mollick at Wharton, point clearly in this direction.
The Power of Role-Play
Our development work with Skill Chamber started with the idea that language students would benefit from speaking practice as homework. Over time, it developed into something much more. A platform where teachers can set AI role-play sessions for students to practise speaking and thinking on the fly, build confidence in oracy, and develop future-proof skills in critical thinking and AI fluency.
Take science. Instead of slogging through another worksheet on photosynthesis, imagine a student stepping into a mock town hall. The AI plays the farmer worried about crops, the councillor fretting about budgets, the teenager angry about the planet’s future. The student has to explain climate data in plain English, defend their position, and adapt on the fly.
Or literature. Rather than another essay on Hamlet, why not cast the student as Hamlet’s therapist? The AI takes the role of the troubled prince, pacing and muttering about ghosts and revenge. The task is to draw him out, ask the right questions, and talk him down from the brink.
Or business. Instead of another tidy case study, picture a student pitching to a panel of AI investors. One demands hard numbers, another interrupts with awkward questions, a third pushes for a social angle. The student has to steady their voice, adjust their delivery, and try to win the room.
Make Homework Meaningful Again
The reason role-play works is simple. Practice makes learning stick. Psychologists have known this for over a century. Ebbinghaus’ forgetting curve showed how quickly we lose what we don’t actively use.
Old homework was about review.
Read it, cram it, forget it.
AI homework can be about rehearsal.
Try it, repeat it, own it.
Instead of producing yet another essay destined for the bin, students put their knowledge to work in a conversation, a negotiation, a pitch. It’s not only more engaging. It’s more effective.
These scenarios aren’t just about photosynthesis, Hamlet, or supply and demand. They prepare students for the world they’re growing up into. One where adaptability, communication, and emotional intelligence matter as much as raw knowledge.
If the old homework trained clerks, the new homework should train collaborators, problem-solvers, and leaders.
That’s the vision behind Skill Chamber. A place where skills come to life.
Conclusion
AI has exposed the fault lines in traditional homework. When answers are cheap, copying becomes tempting, challenge disappears, and motivation quietly collapses. That’s the risk educators are rightly worried about.
But it’s also an opportunity.
When homework shifts from producing answers to practising thinking, those risks fall away. You can’t copy and paste your way through a role-play, a conversation, or a live problem. Students have to engage, respond, and adapt. The work becomes visible again.
Used well, AI doesn’t remove difficulty. It helps pitch it. Each student can be stretched at the right level, receive feedback in the moment, and build confidence through use rather than recall.
That’s what meaningful homework looks like in an AI-rich world. Not a workaround, but a redesign. One that puts practice, judgement, and human interaction back at the centre.
Teachers as Coaches
Next week, we’ll examine the role of the teacher in a world with AI.
Used well, AI offers real opportunities to free teachers from the grind of marking essays. That time can be reinvested where it matters most: high-value tuition and human connection. Teachers step into the role of coach. They set the challenges, cast the roles, and help students make sense of what just happened.
AI can add intensity and realism, but the learning sticks because a human teacher ties it back to meaning.
First in a three-part series exploring how AI role-play could transform education
I don’t know how school was for you.
For me, life didn’t really begin until after university.
After school, I felt top heavy, my head full of information and my body out of place in the real world. I knew how to write essays, pass exams, and play the academic game. I lacked practical know-how or survival instincts.
Instead of stepping neatly onto the first rung of a career ladder, I moved to Spain. That choice was personal and unconventional, but over time I realised something else was shifting too.
Later, when I started my own businesses, I learned more in a few chaotic summers in hospitality than I ever did in Business Studies A-level. Marketing. Leadership. People skills. Making decisions on the fly. Handling pressure when things go wrong.
Everyone’s path is different. But increasingly, employers are looking for capability and credibility, not just qualifications. Experience, not exposure. Judgement (nous) not recall.
And I can’t help thinking the education system could do more to prepare students. Not just for the world as it was, but for the world that is to come.
Educators are not standing still. Many schools are actively strengthening oracy, collaboration, and deeper thinking. Core subjects like language and science are being taught with more care, more awareness than in the past.
The tension isn’t about ambition or effort. It’s structural.
On one side, schools rely on quantitative exercises that are quick to assign, easy to grade, and defensible under auditing frameworks. On the other, there is written work designed to encourage reasoning and expression, but which demands significant teacher time to read, assess, and respond to properly.
Both approaches exist for good reasons. Both affect workload and are prescribed by governing bodies.
Some forward-thinking schools are beginning to experiment with AI-assisted evaluation to ease this burden. In Europe, however, regulatory and ethical considerations rightly slow widespread adoption.
The question remains are students assessed so they can benefit from the feedback OR so the school can monitor progress and the educational authorities can concoct league tables?
From the student’s perspective, this distinction matters.
Feedback frequently arrives days or weeks later, when the work is already a distant memory. The moment of effort has passed. The opportunity to adjust thinking is gone.
At the same time, students are increasingly aware that many homework tasks can now be completed by generating competent answers with AI tools. When that happens, they get the job done but do not learn from the effort.
This is the paradox modern education is navigating.
How can we address this growing mismatch between how learning happens and how success is measured.
Homework sits right at the centre of this tension.
In theory, it exists to reinforce learning. In practice, it often does the opposite.
I remember doing homework on the bus, half-remembering what we’d covered in class. Anything interactive or discussion-based stuck. Course material I read in text books felt harder to access. Nevertheless without much effort I got good grades.
I learnt how to please the system.
Homework rewarded repetition and form. Essays can be crafted to tick the boxes. It's generally quite easy to give them the answers they are looking for. I was quietly compliant.
Homework was built on an industrial-age model that treated students like individual workers completing isolated tasks. A legacy of the old three Rs, still echoing through modern education.
The flaws are obvious when you step back.
Memorisation is prioritised in a world where facts are instantly searchable.
Much of it is lonely work, despite the fact that most real-world problem solving is deeply collaborative.
Paraphrasing is rewarded, as students learn to regurgitate coursework rather than form original judgements.
Feedback arrives late or in generic form, long after the moment of thinking has passed.
Looked at this way, it’s easy to see why so many students disengage. Not because they don’t care, but because the task no longer maps cleanly onto the skills the real world demands, or onto how humans are actually motivated to learn.
Even the best educators know this.
In class, great teachers create discussions, simulations, and moments of genuine engagement. But once students leave the room, practice usually collapses back into worksheets, essays, and solitary tasks.
You can’t practise empathy on a worksheet.
You can’t build negotiation skills through multiple choice.
And you don’t develop confidence under pressure by writing alone at your bedroom desk.
No athlete improves by watching others perform. No musician gets better by rereading sheet music.
They practise.
For those who worry that the use of AI will make students forget how to think, it’s worth pausing here. The risk does not lie in the technology itself, but in how it is used.
Used poorly, AI can flatten learning into shortcut-seeking and surface-level completion. Used well, it can do the opposite.
AI can act as a sparring partner that challenges ideas rather than supplying answers. A collaborator in co-creation rather than a shortcut to completion. A mentor that supports skill development through guided practice, personalised feedback, and repetition that leads to mastery.
Seen this way, the opportunity becomes clear.
After the widespread distribution of knowledge through textbooks, online courses, and video classes, the next frontier in education is not access to information, but the scaling of practice.
This isn’t about blaming teachers. Most are doing extraordinary work inside a system that hasn’t kept pace with change.
Reading and writing still matter. Knowledge still matters.
But if homework is going to earn its place in modern education, it needs to evolve. Away from passive review. Towards active rehearsal.
Towards tasks that feel real. Decisions that carry weight. Conversations that require judgement, not recall.
The question is no longer whether students should practise real-world skills outside the classroom.
It’s how we make that practice accessible, scalable, and meaningful.
That’s the question the next part of this series will explore.
Because the future of homework isn’t about doing more work.
It’s about doing the right kind.
Next:
Part 2 proposes an update to homework that gives students the chance to consolidate skills by interacting with AI whilst keeping teachers in the centre of the process.
Welcome to WordPress. This is your first post. Edit or delete it, then start writing!
