What's at Stake? Young People's Take on AI in their Education Futures
Domain
Education, EdTech, AI in Education
Problem
Representation, agency, AI literacy, inclusive design
Methods
Participatory design, speculative design, zine-making, co-design
AI is arriving in classrooms faster than anyone has asked students what they think — so we worked with 22 young people across three UK secondary schools, using zine-making, speculative design, and hands-on AI experimentation to surface what students actually want from it, and found they are precise, critical, and largely unheard.
The Problem
AI is moving into classrooms faster than anyone has asked students what they think. Policymakers write frameworks. Companies build tools. Teachers implement guidance. And the people whose learning, data, and futures are most directly at stake are rarely in the room when decisions are made. We set out to change that — and to produce findings that could directly inform the designers, developers, and decision-makers shaping these tools.
Who we worked with
An arts organisation, a digital skills organisation and local secondary schools. We collaborated with 22 young people aged 13 to 18 across three UK secondary schools, including a school for students with additional needs. We deliberately included students who are routinely excluded from standard research — including minimally verbal participants whose experiences produced some of the most significant findings of the project.
How we did it
- 01
Digital timelines
Mapping participants' existing relationships with technology.
- 02
Fictional AI company advertisements
Exploring what responsible and irresponsible AI looks like.
- 03
Speculative design sessions
Imagining AI-powered schools of the future.
- 04
Hands-on experimentation
Probing ChatGPT, Microsoft Copilot and Midjourney for bias, accuracy and limitations in real time.
- 05
Zine-making
Students documented their concerns, hopes and demands for policymakers in a collaboratively produced publication — giving voice to participants who found conventional workshop formats harder to access.
What we found
Finding 01
RepresentationAI as a broken mirror
When one student used an AI image generator to create a picture of himself and his best friend, the tool repeatedly produced two white children. He hit recreate again and again, then disengaged entirely. Across all three schools, students described AI as reinforcing stereotypes rather than reflecting reality.
If you are building AI for education
We can help you design and run inclusive research with the students your current process is missing, before those gaps become embedded in your product.
If you are deploying AI in educational settings
We can work with you and your students to understand what is and is not working, and what needs to change before wider rollout.
Finding 02
TransparencyConfident outputs without traceable sources
One group spent hours trying to get an AI tool to generate a historically accurate image. They concluded not that AI was useless, but that its opacity about its own limitations was itself the problem. Students were consistently more comfortable with tools that acknowledged uncertainty than with tools that produced polished-sounding answers they could not verify.
If you are building AI for education
We can research how your specific users experience the gap between AI confidence and AI accuracy, and translate those findings into concrete design direction.
If you are deploying AI in educational settings
We can help you understand that gap through research with your own students, and design the support structures that make deployment genuinely effective.
Finding 03
AgencyStudents want a say, not just a seat
One student created a satirical AI company advertisement with the slogan 'Unlimited: Always Listening, Always Watching' — a sharp critique of AI surveillance. Across all three schools, students expressed clear concern about who controls their data, wanted genuine opt-out options, and — most consistently — did not want AI imposed on them. They wanted to co-create its implementation.
If you are building AI for education
We can design and facilitate co-design programmes that bring young people into your product development as genuine collaborators, not just test subjects.
If you are deploying AI in educational settings
We can design and run participatory research with your students that generates the evidence base you need and gives young people a genuine stake in the outcome.
“AI should be like a teacher's assistant, not the teacher. And it should actually know who we are.”
— Student, age 15
Working with Forth Story
If you are building or deploying AI tools for education and want research that reaches the students your standard processes miss — get in touch. We worked with a wide project team to deliver this work.