AI Failure Intelligence — Real-World AI Risk Dataset

5,950+ structured real-world AI failure cases — severity scored, root cause analyzed, and updated daily. Built for researchers, red-teamers, and enterprise security teams. Each case includes:- Failure Type (14 categories: Security, Hallucination, Autonomous Agent Risk, Alignment Failure, Bias, and more)- Severity Score (0-100)- Display Priority (HIGH/MEDIUM/LOW)- Industry Coverage- Case Type (REAL/RESEARCH/HYBRID) This is a FREE sample of 50 cases.Full dataset (5,500+ cases, 20 columns): https://aifi-project.com