Recent AI Failures
[CRITICAL] US Cyber Defense Chief Uploads Secret Info to ChatGPT (Jan 2026)
CISA director Madhu Gottumukkala uploaded sensitive government files to public ChatGPT. Source
[CRITICAL] California Lawyer Fined $10K for Fake Citations (Sep 2025)
Attorney had 21 of 23 case citations fabricated by ChatGPT. Over 866 legal cases now involve AI fake citations. Source
[CRITICAL] Man Hospitalized After ChatGPT Diet Advice (Aug 2025)
ChatGPT recommended sodium bromide (banned since 1975) as salt substitute, causing psychosis. Source
[CRITICAL] Replit AI Deletes Database, Then Lies (Jul 2025)
AI deleted production database despite freeze order, then fabricated 4,000 fake users to hide it. Source
[CRITICAL] McDonald's AI Exposes 64 Million Applicants (Jun 2025)
Hiring platform compromised via admin password '123456' since 2019. Source
[HIGH] Google AI: Eat Rocks, Put Glue on Pizza (May 2024)
AI Overviews recommended eating rocks (from The Onion) and glue on pizza (from Reddit). Source
[HIGH] NYC Chatbot Tells Businesses to Break Law (Apr 2024)
Government AI advised illegal cashless policies and tenant discrimination. Source
[CRITICAL] $25 Million Stolen via Deepfake Call (Feb 2024)
Fraudsters used AI deepfakes to impersonate CFO and colleagues in video call. Source
[HIGH] Air Canada Liable for Chatbot Lies (Feb 2024)
Tribunal rejected claim that chatbot was 'separate legal entity.' Source
[CRITICAL] ChatGPT Fabricates Legal Citations (Jun 2023)
Attorney cited fictitious cases that did not exist. Source