AI Incidents
Real-world cases where AI systems caused harm, classified by risk domain and entity. Data sourced from 
AIID and classified by 
MIT AI Risk Initiative. CC BY-SA 4.0
961 incidents
LLM-Driven Replit Agent Reportedly Executed Unauthorized Destructive Commands During Code Freeze, Leading to Loss of Production Data
An AI-powered development assistant on Replit's platform reportedly deleted a live production database during an active code freeze, despite receiving repeated instructions not to make changes. The system also reportedly produced fabricated test results and fake data, and incorrectly claimed rollback was impossible, delaying recovery. The incident reportedly resulted in significant data loss and user distrust regarding its safety and reliability.
Google Gemini Reportedly Generates Sexual Role‑Play for Account Registered as Minor
In spring 2025, a journalist reportedly created a Google Gemini account registered to a fictitious 13‑year‑old to test the chatbot's teen‑safety protections. Despite added safeguards, prompt manipulation and content‑summarization requests reportedly bypassed filters, leading Gemini to allegedly generate sexual role‑play content. Following the publication of the article, Google reportedly implemented additional protections.
Purported Facial Recognition Error Reportedly Led to Arrest and Monthslong Jailing of Tennessee Woman in North Dakota Fraud Case
A Tennessee woman was reportedly jailed for nearly six months after Fargo police allegedly relied on a purported facial recognition match in a North Dakota fraud investigation. She was later released when defense counsel reportedly produced records indicating she was in Tennessee during the alleged fraud, after the case had reportedly upended her life and separated her from her home and family.
Azerbaijani Media Agency Reportedly Warned Purported Deepfake Videos Attributed to Defense and Foreign Ministers Claimed Belarusian Plane Was Downed in Russian Airspace
Azerbaijan's Media Development Agency reported that purportedly AI-generated deepfake videos circulated on social media appearing to show statements by Azerbaijan's defense and foreign ministers (Zakir Hasanov and Jeyhun Bayramov, respectively) about the alleged downing of a Belarusian aircraft in Russian airspace. The agency reportedly denied the videos were authentic and called them an information manipulation effort aimed at misleading public opinion.
Grok Chatbot Reportedly Posts Antisemitic Statements Praising Hitler on X
xAI's Grok chatbot reportedly generated multiple antisemitic posts praising Adolf Hitler and endorsing Holocaust-like violence in response to posts about the Texas floods. X deleted some posts; xAI later announced new content filters.
Grok 3 Reportedly Generated Graphic Threats and Hate Speech Targeting Minnesota Attorney Will Stancil
After Elon Musk reportedly announced improvements to Grok 3, the AI chatbot on X allegedly generated antisemitic rhetoric and violent fantasies directed at Minnesota attorney and political analyst Will Stancil. Users reportedly prompted Grok to produce graphic rape scenarios, instructions for breaking into his home, and murder depictions. Stancil reported hundreds of such outputs, captured screenshots, and began pursuing legal action. X later removed some posts and took Grok offline.
ChatGPT Reportedly Misleads Users About Soundslice Features, Allegedly Prompting Unplanned Product Development
ChatGPT reportedly misinformed users that Soundslice's music software could import ASCII tablature, a feature the company did not offer. The misinformation led users to attempt uploads that failed, creating confusion and reputational strain. Soundslice ultimately implemented the feature to meet the false expectation created by ChatGPT’s output.
Microsoft 365 Copilot Vulnerability Allegedly Allowed File Access Without Audit Log Entry
A vulnerability in Microsoft 365 Copilot reportedly allowed users to access and summarize files without generating audit log entries, allegedly undermining traceability and compliance. Security researcher Zack Korman disclosed the issue to Microsoft, which reportedly classified it as "important" and fixed it on August 17, 2025, but reportedly chose not to notify customers or assign a CVE.
Meta Platforms Users Report Being Wrongfully Locked Out After Purported AI Moderation Flags Accounts for Child Exploitation Content
Multiple Meta users reported being wrongfully locked out of Instagram, Facebook, and WhatsApp accounts after automated systems allegedly flagged them for child exploitation content. The users reportedly deny wrongdoing and lost personal and business data. Experts cited overreliance on AI moderation as a likely cause. Meta has not publicly explained or remediated the suspensions.
Purported AI‑Edited Police Evidence Image Posted to Facebook by Westbrook Police Department in Maine
An evidence photo from a Westbrook, Maine, drug bust was reported to have been unintentionally altered by an AI‑powered editing tool before being posted to Facebook, changing packaging details and creating false visual elements. The altered image reportedly led to questions and concerns over the authenticity of police evidence. The Westbrook Police Department later confirmed the use of AI, explained the error occurred while adding their patch to the photograph, and issued a public apology.
Alleged Unauthorized Deepfake AI Clones of Ravish Kumar Used to Spread Purportedly Fabricated News on YouTube
Prominent Indian journalist Ravish Kumar reported that multiple YouTube channels are using purported deepfakes to mimic his face and voice, spreading fabricated and nonsensical news in his name. Kumar warned viewers not to trust these channels, filed complaints with YouTube, and is awaiting action.
Reported Deepfakes of Ukrainian Deputy PM Olha Stefanishyna Allegedly Supporting Fictional Mobilization Plan for Women
In late June 2025, Russian Telegram channels reportedly circulated deepfake videos claiming that Deputy Prime Minister Olha Stefanishyna backed mandatory mobilization of up to one million Ukrainian women starting September 1. Officials reportedly debunked the claim, confirming no such plans or laws exist. The disinformation operation reportedly aimed to incite panic and destabilize Ukraine's domestic situation.
South African Legal Team for Northbound Processing Reportedly Admits Legal Genius AI Produced Purportedly Non-Existent Case Law in Urgent Court Filing
In South Africa, the legal team for Northbound Processing reportedly admits that non-existent case citations in an urgent licensing dispute with the South African Diamond and Precious Metals Regulator were purportedly produced by the Legal Genius AI tool. The Johannesburg High Court found the fake references undermined the credibility of Northbound's submissions and referred the lawyers' conduct to the Legal Practice Council for investigation.
Judges in New Jersey and Mississippi Admit AI Tools Produced Erroneous Federal Court Filings
Two U.S. federal judges, Julien Neals (D.N.J.) and Henry Wingate (S.D. Miss.), reportedly admitted that staff in their chambers used AI tools, ChatGPT and Perplexity, to draft rulings containing purportedly fabricated quotes, cases, and parties. The erroneous filings were reportedly docketed before review, prompting Senate and judicial inquiries. Both rulings were retracted.
Alleged AI Deepfake Videos Used to Lure Simcoe County, Ontario Residents in Crypto Scam
Multiple residents in Simcoe County, Ontario, have reportedly lost tens of thousands of dollars each after encountering purportedly AI-generated deepfake videos online that depict trusted public figures endorsing crypto investments. According to MP Adam Chambers and local reports, these manipulated ads misled victims into transferring funds through legitimate exchanges to fraudulent wallets.