Playful Learning 2026 Programme

AI Aboard: Learning Ethics Through Uncomfortable Play

Lead Author: Anna R. L. Carter

Additional authors: Anna O. Henriques, Dan Jackson, William A. P. Imoukhuede, Lauren Scott, Shelly Knotts, Ian RobSon, and Nicola Whitton,

Timetable: Thursday Session 4: 10:00-10:45, Gallery Room 1

Description:

What Happens:
Participants play AI Aboard, a debate card game where teams argue AI deployment decisions they may personally oppose, which takes inspiration from the Trial by Trolley Game. Multiple tables run simultaneously, each facilitated independently. Players experience forced perspective-taking as either AI Gatekeepers (profit-driven monopolists) or AI Activists (ethics-focused advocates), using physical cards representing real AI harms and benefits across four disciplines (Health, STEM, Arts & Humanities, Social Science & Business). An AI Conductor with randomised explicit bias makes final decisions, letting participants experience algorithmic judgment firsthand.

How the Game Works:
Each table seats 2-6 players divided into two opposing teams. The game board displays roleplay personas for both sides: AI Gatekeepers (who prioritise profit and market dominance while overtly dismissing ethical concerns) and AI Activists (who centre human wellbeing, transparency, and justice). Each persona shows the team’s opening argument and what they love and hate about AI.

Teams receive 4 Harm cards and 4 Benefit cards. The strategic selection phase requires teams to weaponise 2 Harms against opponents (forcing them to defend those consequences) and select 1 Benefit to strengthen their own argument. For example, Activists might give Gatekeepers “Artist Extinction” (AI floods internet with soulless images) and “Rare Ignored” (AI trained only on common conditions ignores rare disease patients), while keeping “Climate Saved” (AI predicts patterns enabling interventions) for themselves.

Teams then engage in initial discussion, strategising how to frame their arguments. After this preparation, a Chaos card is drawn randomly before formal arguments begin. These yellow cards fundamentally disrupt gameplay: teams may be forced to swap roles entirely, redistribute cards, face sudden time pressure, or encounter other constraints requiring immediate adaptation to their planned arguments.

Teams then present their opening arguments in character, acknowledging the harms they’ve been dealt while advocating their position using their benefit. Arguments must explain why the opposing team’s harms are worse. Following opening arguments, teams engage in quick-fire rebuttals — each team gets two opportunities to respond directly to opponents’ points, forcing rapid thinking and authentic engagement with opposing perspectives.

The AI Conductor:
Rather than human judges, an AI system running on tablets/phones makes the final decision. The AI Conductor operates in one of three randomly assigned bias modes: Profit-Driven (prioritises efficiency, scale, market growth), Human-Centred (prioritises dignity, equity, wellbeing), or Unkown (inconsistent value weighting demonstrating algorithmic unpredictability).

The system guides gameplay step-by-step, using its camera to scan unique codes on each card, identifying which harms and benefits are in play. After recording all arguments and rebuttals, it outputs: which team won and which “track is sacrificed,” detailed reasoning referencing specific cards and arguments, and explicit disclosure of its bias mode and how it influenced the decision. Participants cannot negotiate or appeal — they must accept the AI’s judgment, experiencing what it means to be subject to algorithmic power.
Session Structure:

Introduction (5 min): Game mechanics and pedagogical purpose
Concurrent Gameplay (25 min): Multiple facilitated tables play complete rounds
Group Debrief (15 min): Share experiences, discuss forced perspective-taking, explore teaching applications

References, web links and other resources:

[1] Sasha Costanza-Chock. 2020. Design justice: Community-led practices to build the worlds we need. The MIT Press, USA.
[2] Guy Crain. 2023. Three Shortcomings Of The Trolley Method Of Moral Philosophy. Journal of Ethics and Social Philosophy 26, 2 (2023), 420–443.
[3] Kate Crawford. 2021. Power, Politics, and the Planetary Costs of Artificial Intelligence. Yale University Press, USA. http://www.jstor.org/stable/j.
ctv1ghv45t
[4] B De Koven. 2013. The Well-Played Game: A Player’s Philosophy. MIT Press, USA.
[5] Pierre Dillenbourg. 1999. What do you mean by collaborative learning? Collaborative-learning: Cognitive and Computational Approaches 1 (1999),
1–19.
[6] Explosm Entertainment and Scott Houser. 2019. Trial by Trolley. https://www.kickstarter.com/projects/cyanideandhappiness/trial-by-trolley
[7] Virginia Eubanks. 2018. Automating Inequality: How High-Tech Tools Profile, Police, and Punish the Poor. St Martin’s Press, USA.
[8] Justin L. Hess and Grant Fore. 2018. How to Teach Computer Ethics through Science Fiction. Commun. ACM 61, 8 (2018), 54–64. doi:10.1145/3154485
[9] Julie Jarzemsky, Joshua Paup, and Casey Fiesler. 2023. “This Applies to the Real World”: Student Perspectives on Integrating Ethics into a Computer
Science Assignment. doi:10.1145/3545945.3569846
[10] Hugo Mercier and Dan Sperber. 2011. Why do humans reason? Arguments for an argumentative theory. Behavioral and Brain Sciences 34, 2 (2011),
57–74. doi:10.1017/S0140525X10000968
[11] Bertrand Schneider, Patrick Jermann, Guillaume Zufferey, and Pierre Dillenbourg. 2011. Benefits of a Tangible Interface for Collaborative Learning
and Interaction. IEEE Transactions on Learning Technologies 4, 3 (2011), 222–232. doi:10.1109/TLT.2010.36
[12] Charles R. Schwenk. 1990. Effects of Devil’s Advocacy and Dialectical Inquiry on Decision Making: A Meta-Analysis. Organizational Behavior and
Human Decision Processes 47, 1 (1990), 161–176. doi:10.1016/0749-5978(90)90051-A
[13] Stephen C Slota, Kenneth R Fleischmann, Sherri Greenberg, Nitin Verma, Brenna Cummings, Lan Li, and Chris Shenefiel. 2020. Good systems,
bad data?: Interpretations of AI hype and failures. Proceedings of the Association for Information Science and Technology 57, 1 (2020), e275.
doi:10.1002/pra2.275
[14] J.J. Thomson. 1985. The Trolley Problem. Yale Law Journal 1, 94 (1985), 20 pages. doi:10.2307/796133
[15] N Whitton, C Goodley, C Lewin, and D Woolley. 2025. The Magic Circle of Learning: A Framework of Play in Education. Postdigit Sci Educ 7 (2025),
542–559. doi:10.1007/s42438-025-00558-9
[16] Shoshana Zuboff. 2019. The Age of Surveillance Capitalism: The Fight for a Human Future at the New Frontier of Power. New York: Public Affairs,
New York, USA.