Synopsis
This workshop replaces compliance checklists with an integrity-focused escape room approach that emphasizes collaborative ethical reasoning and context-based decision-making. It also demonstrates responsible use of generative AI to convert existing training materials into experiential learning across professional fields.
Abstract
As artificial intelligence (AI) systems become increasingly embedded in decision-making across education and other professional domains, there is growing pressure to address AI ethics in ways that extend beyond policy awareness, technical competence, or compliance-driven training. In teacher education in particular, ethical frameworks, competency models, and institutional guidelines provide important reference points (UNESCO, 2023), yet prior scholarship suggests that principle-based approaches alone often fail to support practitioners in navigating the situated, value-laden decisions they encounter in everyday practice (Floridi, 2018; Morley et al., 2020). Educators may be able to articulate ethical principles in theory yet still struggle to operationalise them when faced with time pressure, institutional constraints, opaque AI systems, and competing stakeholder expectations.
This workshop proposes and demonstrates an integrity-led escape room pedagogy for AI ethics training that is domain-flexible by design, with teacher education and professional learning used as the primary instantiation. The approach is reimagined from commonly taught AI ethics training content, including distinctions between ethics and morality, means versus ends, responsibility and accountability, bias and fairness, transparency, privacy, and trust.
The workshop is structured in three interconnected stages. In the first stage, participants are briefly oriented to selected AI ethics concepts through short framing prompts drawn from existing training materials used in teacher education and professional development contexts. This stage does not involve extended lecturing; instead, it establishes shared conceptual anchors and ethical tensions that will later be explored experientially. Participants are encouraged to consider how ethical decision-making in AI-enabled environments is shaped not only by outcomes, but by the processes through which decisions are made, a concern that resonates across education, healthcare, and other professional domains.
In the second stage, participants work in small groups to complete an AI ethics escape room. The escape room consists of a series of “locked” stations, each aligned to a core ethical theme such as means versus ends, bias and fairness, privacy and surveillance, accountability, or trust in AI-mediated decision-making. At each station, groups are presented with a fictionalised but realistic scenario involving AI use. While scenarios are demonstrated through teaching and institutional contexts, their structure is intentionally adaptable, allowing the same ethical tensions to be explored in domains such as healthcare, public services, or organisational decision-making. To progress, groups must collectively discuss the dilemma, articulate ethical risks, identify affected stakeholders, and justify their decision-making process There are no singular correct answers; progression is based on the quality of ethical reasoning, supported through active facilitation that prompts reflection and surfaces underlying assumptions.
In the third stage, the workshop shifts to a meta-pedagogical focus, demonstrating how generative AI tools such as ChatGPT can be used responsibly as design partners to convert existing AI ethics training into experiential learning formats. Participants are guided to use conceptual prompting and abstraction rather than proprietary or sensitive materials, foregrounding data minimisation and human-led ethical framing. This positions generative AI as a bounded pedagogical aid that supports design without displacing human judgment.
The pedagogical design of the workshop is grounded in constructivist learning theory, which emphasises ethical understanding through dialogue, social interaction, and negotiated meaning (Vygotsky, 1978), and experiential learning theory, which frames learning as a cycle of concrete experience, reflective observation, conceptualisation, and active experimentation (Kolb, 1984). Ethical decision-making is further informed by moral reasoning perspectives that prioritise justification, responsibility, and reflection overrule compliance or outcome-based optimisation (Quinn, 2015). These theoretical foundations are deliberately operationalised through the escape room format, which legitimises uncertainty, disagreement, and ethical discomfort as productive learning conditions across professional learning contexts.
The workshop design was informed by feedback from a teacher incubator programme in which participants highlighted the value of interactive discussions, informal peer exchange, exposure to new AI tools, and structured yet welcoming learning environments. Participants reported that unplanned conversations often triggered deeper ethical reflection and professional insight. These insights directly shaped the escape room methodology, reinforcing the importance of social learning, facilitation, and reflective debriefing in AI ethics training more broadly.
Following the escape room, participants engage in structured reflection and feedback activities focused on the pedagogical effectiveness, ethical depth, and transferability of the method. The workshop adopts a design-based orientation, using participant feedback to iteratively refine the escape room pedagogy across domains. By combining experiential ethics education with transparent demonstrations of responsible generative AI use, the workshop contributes both a pedagogical model and a replicable design approach for AI ethics training in education, healthcare, and other professional settings.
This session contributes to scholarship on AI ethics education and professional learning by offering a theoretically grounded, practice-oriented methodology that moves beyond checklist ethics toward reflective, trust-building approaches. It demonstrates how generative AI can be used ethically to support pedagogical design rather than replace human judgment and provides participants with concrete strategies for transforming existing training materials into engaged, integrity-led learning experiences across domains.
Target Audience
Teacher educators and academic developers
Professionals involved in AI-enabled decision-making (education, healthcare, public services)
Researchers in AI ethics, academic integrity, trust, and the scholarship of teaching and learning
Curriculum designers, trainers, and policy implementers responsible for ethical AI integration
Key Takeaways
A pedagogically rigorous model for teaching AI ethics through escape room methodologies that foreground ethical reasoning over compliance.
Practical strategies for responsible use of generative AI to redesign existing training into engaged learning experiences without compromising copyright or data privacy.
A transferable framework for integrity-led, experiential AI ethics education across education and other professional learning contexts.
Keywords
AI Ethics, GenAI, Education, Integrity Training, AI-enabled decision-making, pedagogy
References
Floridi, L. (2018). Soft ethics and the governance of the digital. Philosophy & Technology, 31(1), 1–8. https://doi.org/10.1007/s13347-018-0303-9
Kolb, D. A. (1984). Experiential learning: Experience as the source of learning and development. Prentice Hall.
Morley, J., Floridi, L., Kinsey, L. et al. (2020) From What to How: An Initial Review of Publicly Available AI Ethics Tools, Methods and Research to Translate Principles into Practices. Sci Eng Ethics 26, 2141–2168 (2020). https://doi.org/10.1007/s11948-019-00165-5
Quinn, M. J. (2015). Ethics for the information age (7th ed.). Pearson.
UNESCO. (2023). AI competency framework for teachers. UNESCO Publishing. https://www.unesco.org/en/articles/ai-competency-framework-teachers
Vygotsky, L. S. (1978). Mind in society: The development of higher psychological processes. Harvard University Press.