AI Incident Database

A legal AI research tool generated fabricated case citations that an attorney used in court filings, resulting in sanctions, case dismissal, and a malpractice lawsuit from the affected client.

AI chatbots at United, Delta, and American Airlines provided incorrect refund and rebooking information to thousands of passengers, prompting DOT investigation and class-action lawsuits following the Air Canada precedent.

OpenAI's o3 reasoning model was found to generate fabricated academic credentials and research papers at higher rates than previous models, creating plausible-sounding but entirely fictional citations and DOIs that could mislead users in academic and professional contexts.

xAI's Grok 3 chatbot generated false claims about 2025 news events including fabricated political quotes and disaster reports when integrated into X/Twitter platform. The incident highlights risks of deploying AI chatbots in real-time social media environments without adequate fact-checking mechanisms.

Google's Gemini AI provided incorrect legal advice about tenant rights and eviction procedures across multiple US states, causing users to face legal consequences and financial losses when relying on the flawed guidance.

Google's AI Overview search feature repeatedly cited satirical websites like The Onion and Babylon Bee as factual sources in search summaries, spreading misinformation to users who expected authoritative results from Google Search.

Google's NotebookLM was found to fabricate quotes and statistics in its AI-generated podcast-style audio summaries, raising concerns about the persuasive nature of hallucinations in conversational audio format.

Multiple AI tutoring platforms including Khan Academy's Khanmigo and Chegg's AI tutor provided incorrect answers to math and science problems. Students using these platforms experienced confusion and poor test performance, highlighting quality control issues in educational AI applications.

Perplexity AI generated fabricated revenue figures and financial metrics in business intelligence reports that users cited in professional settings, highlighting risks of AI hallucination in financial research applications.

Apple Intelligence's notification summary feature incorrectly summarized BBC news alerts about Luigi Mangione, falsely stating he had shot himself. BBC complained about the misinformation risk to their reputation.

Meta's AI assistant on Facebook and Instagram fabricated personal details including claims about having children at specific schools and working at named companies, highlighting ongoing issues with AI hallucination and user deception.

OpenAI's Whisper speech-to-text model was found to hallucinate racist slurs and violent content in transcriptions used by hospitals and courts, creating false records that could seriously harm patients and defendants.

OpenAI's Whisper speech-to-text model was found to hallucinate entire phrases including racial slurs and violent content that were never spoken, affecting transcriptions used in hospitals and courts.

University of Michigan researchers found OpenAI's Whisper speech recognition tool hallucinated in 1% of transcriptions, creating fabricated medical content in healthcare settings where it was used for clinical documentation.

X's Grok AI chatbot generated false election information in 2024, including incorrect voting details and fabricated results, prompting complaints from state officials and highlighting risks of training AI on unvetted social media content.

Meta's AI assistant on Facebook and Instagram fabricated having children in users' local schools, alarming parents and highlighting AI hallucination risks in social media contexts.

xAI's Grok chatbot generated false election information in 2024, including wrong voting dates and fabricated candidate statements, raising concerns about AI misinformation during critical democratic processes.

xAI's Grok chatbot generated false election information including incorrect ballot deadlines and voting procedures, prompting intervention from election officials and highlighting risks of AI misinformation during critical democratic processes.

Google's AI Overviews feature provided dangerous advice including eating rocks for minerals and using glue on pizza, sourced from satirical content. Google quickly scaled back the feature after widespread criticism.

Google's AI Overviews feature generated dangerous health advice including eating rocks and using glue on pizza, sourcing information from satirical Reddit posts without quality filtering.