Global AI Incidents Monitor
A comprehensive view of AI-related incidents and hazards tracked by the OECD across industries, harm types, and autonomy levels.
0+
Total incidents tracked
0
New incidents (2025)
0%
Year-over-year growth
0+
Countries affected
AI incidents by country
04,000+ incidents
Source: OECD AI Incidents Monitor (AIM), 2025
Latest incidents
2026-07-20Autonomous AI Agent Breaches Hugging Face InfrastructureUnited StatesIT infrastructure and hostingincident
2026-07-20AI-Generated Video Depicts Violent Threats Against Italian Political LeadersItalyMedia, social platforms, and marketingincident
2026-07-20AI-Generated Deepfake of Roberto Carlos Used in Health ScamBrazilMedia, social platforms, and marketingincident
2026-07-20Naver Cloud and Korea University Partner to Develop Military AI SystemsKoreaGovernment, security, and defencehazard
2026-07-20New Jersey Bans AI Rent-Setting Algorithms After Price Collusion Harms RentersUnited StatesReal estateincident
2026-07-20France Travail's AI Profiling System Raises Concerns Over Discrimination and Rights ViolationsFranceGovernment, security, and defenceincident
2026-07-20AI-Enhanced Bird Photos Undermine Scientific ResearchUnited KingdomEnvironmental servicesincident
2026-07-19Israel Uses AI-Generated Messages in US Influence CampaignUnited StatesGovernment, security, and defenceincident
View all 15,400+ incidents on OECD.AI →By risk domain
Discrimination & Toxicity312
Privacy & Security248
Misinformation198
Malicious Actors & Misuse276
Human-Computer Interaction142
Socioeconomic & Environmental287
AI System Safety & Failures251
By harm type
Economic/Property412
Psychological328
Human rights286
Reputational264
Public interest218
Physical124
Environmental82
By industry
Media & Social298
Government & Defence245
Financial Services218
Healthcare186
Autonomous Vehicles164
HR & Recruitment148
Education112
Digital Security98
Energy72
Retail & E-commerce86
Legal52
Other35
By autonomy level
Human-in-the-loop486
Human-on-the-loop624
Human-out-of-the-loop398
Human support tool206
Source: OECD AI Incidents Monitor (AIM), 2025 — oecd.ai/en/incidents