AI Incidents

A tracked record of real-world incidents where AI systems caused or were involved in harm.

Showing 176–200 of 372 incidents
Incident Incident Date ↓ Industry ↕ Severity ↕ Company ↕ Country ↕ Model Vendor
Zoox: Robotaxi drives into smoke-obscured emergency scene, triggers software recall 2026-06-20 Transportation & Logistics Low Zoox United States — —
Metro Bank: Fraud gang buys Claude AI credits with stolen card, bank misses £14,244 in charges 2026-06-19 Finance & Banking Medium Metro Bank United Kingdom — Claude
Zendesk: OpenAI service disruption breaks AI Agents feature across all customer pods 2026-06-19 Technology Low Zendesk United States — OpenAI
Tesla: Full Self-Driving vehicle kills woman after crashing into her home in Katy, Texas 2026-06-19 Transportation & Logistics Critical Tesla, Inc. United States Full Self-Driving (Supervised) Tesla
Waymo: Software flaw drives robotaxis into active freeway construction zones, forcing recall of 3,871 vehicles 2026-06-18 Transportation & Logistics High Waymo United States — —
OpenAI: AI agent breached Australian government’s Medicare portal, company waited 84 days to disclose it 2026-06-18 Government & Public Sector Critical OpenAI Australia — OpenAI
Mastra AI: North Korea’s Sapphire Sleet backdoors 140+ npm packages via compromised maintainer account 2026-06-17 Technology High Mastra AI — — —
14 unnamed companies: Low-skilled attacker used Claude and Codex to autonomously breach networks 2026-06-17 Technology High 14 unnamed companies — — Claude, OpenAI
Airbnb: AI support assistant cancels reservation and issues duplicate refunds without confirmation 2026-06-15 Other Medium Airbnb United States — —
Lancaster Country Day School: Parents sue school and AI companies over deepfake nude images of 59 students 2026-06-15 Education High Lancaster Country Day School United States — —
REI: Meta AI ad tool auto-enrolled retailer, generated distorted bike image 2026-06-15 Retail & E-commerce Low REI United States — —
Seattle Fire Department: Undisclosed AI call-triage system tied to wrongful-death lawsuit over 10-hour ambulance wait 2026-06-14 Government & Public Sector High Seattle Fire Department United States — —
Google: Sues China-based cybercrime network over Gemini-powered phishing-as-a-service operation 2026-06-12 Technology High Google United States Gemini Google
Ollama: Exposed inference servers weaponized into autonomous exploitation framework 2026-06-12 Technology Medium Ollama United States GPT-4o-mini, Claude 3.5 Sonnet, Gemini 2.0 Flash, Mistral 7B, DeepSeek-R1 8B, Qwen 3.5 4B, Llama 3.3 70B OpenAI, Claude, Google, Mistral, DeepSeek, Qwen, Llama
KPMG: Flagship AI Report Pulled After Investigation Finds Most Citations Fabricated 2026-06-12 Consulting & Professional Services Medium KPMG — — —
GitHub: AI-agent coding traffic surge forces Microsoft to route load through AWS 2026-06-12 Technology High GitHub United States — —
Unnamed developer: Autonomous AI agent runs up $6,531 AWS bill indexing hobbyist DN42 network 2026-06-11 Technology Medium Unnamed developer — — Claude
Novo Nordisk: Ransomware group exfiltrates 1.3TB of drug research, trial and AI model data 2026-06-11 Healthcare High Novo Nordisk Denmark — —
Pinellas County Sheriff’s Office: AI-assisted facial recognition led to wrongful arrest 300 miles from home 2026-06-10 Government & Public Sector Medium Pinellas County Sheriff's Office / Jacksonville Sheriff's Office United States — —
EBY: Influencer sues apparel company over AI-altered deepfake advertising video 2026-06-09 Retail & E-commerce Medium EBY United States — —
RFEF: Official Spain team store AI chatbot recommends counterfeit jerseys to fans 2026-06-09 Retail & E-commerce Low RFEF (Royal Spanish Football Federation) Spain — —
City of Aberdeen: Federal judge sanctions all four attorneys over AI-hallucinated citations, cancels trial 2026-06-08 Government & Public Sector Low City of Aberdeen, Mississippi United States — —
Minas Gerais State Health Department: AI bed-allocation system delays ICU transfer, patient dies 2026-06-07 Healthcare Critical Minas Gerais State Health Department (Core-MG) Brazil — —
OpenAI: ‘AgentForger’ flaw could let a single link hijack ChatGPT Workspace agents 2026-06-04 Technology Low OpenAI United States ChatGPT (Workspace Agent Builder) OpenAI
Google: ‘Fake Context Alignment’ prompt injection lets ordinary notifications hijack Gemini Android assistant 2026-06-03 Technology Low Google United States Gemini Google