AI Education Futures — a Participatory Design Futuring study with the Co-Design Stories Toolkit
Domain
Education, EdTech, Child-centred AI
Problem
Participation, child rights, inclusive design, AI futures
Methods
Participatory design futuring, co-design toolkit, speculative design
Children are almost entirely absent from the design of AI systems that will shape their education — so we put 14 children aged 3 to 10 in charge of designing AI for their own schools using our Co-Design Stories Toolkit and a live roleplay with a fictitious sentient school building, and found they produce sophisticated, values-driven visions that adult-led design processes consistently miss.
The Problem
AI is arriving in classrooms faster than the policy frameworks designed to govern it. Governments are writing guidance. Companies are building products. And children — the people whose education, data, and futures are most directly at stake — are almost entirely absent from the design process. We set out to change that. Not by asking children to evaluate AI systems someone else had designed, but by putting children in charge of designing AI systems themselves — and listening carefully to what they built.
Who we worked with
We worked with 14 children aged 3 to 10 and 8 parents across 7 families, recruited through a Glasgow Science Centre community outreach programme supporting local low-income families. None of the children had received formal AI education at school. This was a deliberate choice. We wanted to hear from children whose understanding of AI was unmediated by curriculum — to understand what children actually think, not what they have been taught to think.
How we did it
- 01
Co-Design Stories Toolkit
A purpose-built participatory research toolkit using a gameboard design sprint format that is adaptable to any context.
- 02
SmartSchool speculative framing
Participants were transported 100 years into the future to design AI for a sentient school building from 2123, roleplayed live by the research team responding to children's ideas in real time.
- 03
Children as lead designers
Children led ideation, client feedback, prototype crafting with physical materials, and final presentations. Adults acted as design assistants. No prior knowledge of AI or design was required.
What we found
Finding 01
Environmentally conscious AI learning futuresChildren designing AI to solve real problems they cared about
Climate and environment were at the forefront of several designs without being prompted. One ten-year-old designed SmartCloudSchool — a system powered by rainwater flowing through turbines, with AI embedded in the school walls to remind students of instructions in science class and give personalised support when the teacher was too busy. He described his design with engineering precision and explained the water cycle from memory. He said he wants to be an engineer. Another child designed RoboVax — a drone-and-robot system that detects rubbish around the school, vacuums what it can, and deploys a claw for larger items, powered by a solar panel. Children were not designing AI for AI's sake.
If you are building AI for educational or public sector contexts
We can design and run research that surfaces what children and young people actually value — so your product reflects genuine needs rather than adult assumptions about what children want from technology.
Finding 02
Immersive and embodied AI learning futuresLearning pulled out of textbooks and into the body
Children imagined AI that pulled learning out of textbooks and into the body. Mrs Gloey the Jellyfish Teacher was a wearable ring: press a button, and a holographic book pops out with a jellyfish AI teacher who explains whatever you want to learn about underwater life. SpyGlasses were glasses that answered questions about whatever you looked at — tested on a painting in an art gallery — and gave you a quiz at the end. These designs are not naive. They reflect a sophisticated intuition about personalisation, embodiment, and learning by interest rather than by curriculum. They are also, notably, not screens.
If you are building educational technology or wearable AI
We can run participatory design research with children that generates the kind of design direction that user testing of existing products cannot — because it asks what children would build, not whether they can use what you have already built.
Finding 03
Personal yet socially connected AI learning futuresPersonalisation, connection, and data on children's terms
The most consistent theme across all groups was the combination of personalisation and social connection. HelloSchool would remember every student's name, favourite animals and colours, know when students were tired and give breaks, and sing to children having a bad day — in different voices depending on what you like. SmartPet offered a holographic pet you could walk to school, charge in a smart cage, and that would help you learn and develop a personality over time. One group imagined SmartSchool talking to the Science Centre — buildings communicating with each other about what their students needed to learn. The scale of that imaginary is striking from a nine-year-old. Children also raised data privacy questions without being prompted — envisioning data shared between trusted institutions they already knew, and expressing the intuition that data sharing should serve them, not the system.
If you are building AI products for children, schools, or families
We can research how children conceptualise personalisation, connection, and data in ways that are child-rights compliant and commercially grounded — and translate those findings into design principles your team can act on.
“SmartSchool has a smart cloud — this is the brain of the smart school [...] It also helps teachers in science classes when they forget things. SmartSchool sends reminders.”
— Child, age 10, describing SmartCloudSchool
Working with Forth Story
If you are building or deploying AI for children, education, or public sector contexts and want research that genuinely centres children's voices — get in touch. We worked with a wide project team to deliver this work.