Students tune out abstract ethics lectures. AI ethics education lands better when it's built around dilemmas they'll actually face, not philosophical thought experiments.
Start with a real scenario, not a definition
"Your friend used AI to write 80% of an essay and got a good grade — is that fair?" generates more genuine discussion than opening with a definition of academic integrity.
Cover bias with a concrete example
Show students an AI system that got something demonstrably wrong due to biased training data. Abstract warnings about "AI bias" don't stick; a specific, visible failure does.
Discuss disclosure as a norm, not just a rule
Help students understand why disclosure matters — trust, fairness, and their own long-term skill development — not just that it's required. Rules without reasons get worked around.
Make it an ongoing conversation, not a single unit
AI capabilities change fast enough that a one-time ethics lesson goes stale within a year. Building in periodic check-ins keeps the conversation relevant as tools evolve.