Responding to Generative AI Technologies with Research-through-Design: The Ryelands AI Lab as an Exploratory Study

TL;DR

This paper investigates how Research-through-Design can guide constructionist education about generative AI in a primary school, using the Ryelands AI Lab as a six-week exploratory curriculum. By engaging students directly with a text-to-image model (Stable Diffusion) and iterating learning materials in close collaboration with teachers, the study demonstrates co-development of practical and critical AI competencies and argues for critical responsivity in design. It contributes actionable guidance for constructing constructionist AI curricula and frames RtD as a rapid-response methodology for addressing the socio-technical uncertainties of rapidly evolving AI technologies, with implications for HCI research and education practice. The work highlights the value of long-form, hands-on engagement with real AI tools to foster reflective critique, teacher empowerment, and inclusive conversations about ownership, creativity, and harms in AI-enabled education.

Abstract

Generative AI technologies demand new practical and critical competencies, which call on design to respond to and foster these. We present an exploratory study guided by Research-through-Design, in which we partnered with a primary school to develop a constructionist curriculum centered on students interacting with a generative AI technology. We provide a detailed account of the design of and outputs from the curriculum and learning materials, finding centrally that the reflexive and prolonged `hands-on' approach led to a co-development of students' practical and critical competencies. From the study, we contribute guidance for designing constructionist approaches to generative AI technology education; further arguing to do so with `critical responsivity.' We then discuss how HCI researchers may leverage constructionist strategies in designing interactions with generative AI technologies; and suggest that Research-through-Design can play an important role as a `rapid response methodology' capable of reacting to fast-evolving, disruptive technologies such as generative AI.

Figures (11)

• Figure 1: Overview of the exploratory study phases as they manifested through the RtD methodology in hindsight.
• Figure 2: Image from early stage of co-design process, showing second author (middle) and the year 4 teachers.
• Figure 3: Two versions of the image generation tool used during delivery. Top: In this version used from week 2 onwards, students can input the main text prompt; and optionally specify negative prompt(s), an initial seed, and the context-free guidance scale, and submit to a custom Stable Diffusion API endpoint. Bottom: A version created for reflection on issues such as ownership and creativity: students select from 5 randomly sourced preexisting image generations and their prompts and can 'convert' these to a particular artist's style.
• Figure 4: Exemplary scans of the worksheets for each week which students would keep in folders. Weeks 1 and 2 feature an introductory worksheet first (e.g, "Describe the artworks below" in week 1). Weeks 3 and 4, with the introduction of the "Reimagining Ryelands" lab project, first inspires students (e.g., by prompting them to list all things they can think of relating to their school) and then asks students to log their reimagination experiments. Week 5 first asks students to log their experiments with artist' styles, and then provides a questionnaire for reflection. Week 6 features individual worksheets which show their chosen prospectus contribution and asks them to write a persuasive accompanying text.
• Figure 5: Left: A cropped print file of the 5x1 meter strip showing an introduction to the week's delivery contents and a random selection of 250 images with their respective prompts. Right: Overview of the week strips laid out in the school gymnasium for the exhibition. In the background, three posters can be seen that show the spreads of the Ryelands AI Lab prospectus.