Key figure: 63% of organizations in IBM's 2025 breach study had no AI governance policy, and 97% of AI-breach victims had no proper AI access controls (IBM, Cost of a Data Breach Report 2025).
AI agents 2026 fail in predictable patterns: prompt injection, data exfiltration, hallucinated actions, runaway costs, agents exceeding authority. Bingo turns the pattern library into a game — mark each failure mode as you recognize it.
AI incidents repeat the same patterns: prompt injection, data exfiltration, hallucinated actions, runaway costs, and agents exceeding their authority. Bingo turns the pattern library into a game — mark each failure mode as you recognize it. Players learn the taxonomy of AI failure faster playing than reading, and the card doubles as a review checklist.
AI Incident Bingo is a quiz game built on real AI failure patterns — prompt injection, exfiltration, hallucinations, runaway spend — that teaches the taxonomy of AI incidents through play.
A cybersecurity quiz game in bingo form: 24 real AI-incident types on one card. Mark the ones you've seen this month — first to five in a row wins the world's most depressing bingo.
Every square is a documented incident type from our incident tracker. Click a square to mark it.
Good cybersecurity quiz questions test judgment against real incidents — like the 24 on this card: prompt injection, data poisoning, rogue agent actions. Scenario-based questions beat trivia because they mirror what teams actually face.
Keep it short, repeatable, and relevant: play this bingo card with your team monthly, walk through each marked incident, and pair it with a phishing quiz. Fifteen minutes of real examples beats an annual slide deck.
Yes — the card is free to play, print, and share. No signup, nothing installed.