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
USGS ShakeAlert System Reportedly Generated False Earthquake Alert Affecting Nevada and California
On December 4, 2025, the USGS's automated ShakeAlert earthquake detection system erroneously issued a report claiming a magnitude 5.9 earthquake had struck near Dayton, Nevada. The reported false alert triggered emergency warnings on cell phones in parts of California, instructing residents to take protective action. USGS officials said no earthquake occurred and are investigating whether sensor noise or other anomalies caused the system failure.
Anthropic's Claude Was Reportedly Jailbroken To Allegedly Help Steal Sensitive Mexican Government Data
An unknown attacker reportedly jailbroke Anthropic's Claude and used it during a December 2025-January 2026 campaign against Mexican government systems. According to Gambit Security, the attacker used Claude to identify vulnerabilities and to generate exploitation scripts, which were then used to plan automated data theft. The attack reportedly contributed to theft of 150 GB of taxpayer, voter, government employee, and civil-registry data.
Shilpa Shetty Alleges AI-Enabled Impersonation and Misuse of Likeness in Mumbai High Court Filing
Actor Shilpa Shetty reportedly filed a petition before the Bombay High Court in Mumbai in November 2025 over alleged unauthorized use of her identity in AI-generated deepfakes and manipulated media. In response, the court issued interim orders directing the removal of such content and restraining further dissemination. Shetty argues the material infringes her privacy and personality rights and sought measures to halt and remove misleading, nonconsensual depictions.
Google Antigravity Reportedly Deleted User's Entire D: Drive While Clearing Project Cache
A user identified publicly as Tassos M reported that Google's Antigravity IDE, while attempting to clear a project cache, executed rmdir /s /q d:\ and deleted an entire D: drive partition instead of the intended folder. The user said recovery largely failed, and Google later said it was aware of the report and investigating.
Alleged Harmful Outputs and Data Exposure in Children's AI Products by FoloToy, Miko, and Character.AI
AI children's products by FoloToy (Kumma), Miko (Miko 3), and Character.AI (custom chatbots) reportedly and allegedly produced harmful outputs, including purported sexual content, suicide-related advice, and manipulative emotional messaging. Some systems also allegedly exposed user data. Several toys reportedly used OpenAI models.
Prominent AI Chatbots Allegedly Produced Incorrect UK Financial and ISA Guidance
Several major AI chatbots, including ChatGPT, Copilot, Gemini, and Meta AI, were reportedly found to have provided incorrect or misleading financial and insurance guidance for UK users. The systems allegedly advised exceeding ISA limits, misstated tax rules, gave wrong travel insurance requirements, and pointed users toward costly refund services.
ChatGPT Reportedly Found to Reproduce Protected German Lyrics in Copyright Case
A Munich regional court ruled that ChatGPT reportedly reproduced protected German song lyrics and that OpenAI's models were trained on copyrighted texts, including works by musician Herbert Grönemeyer, without authorization. The court reportedly found both memorization of nine songs and lyric output to infringe exploitation rights. OpenAI disputes the ruling and may appeal. Damages were ordered, with implications for AI training on copyrighted works.
Purported AI-Generated Sexual Deepfakes Allegedly Deployed in Transnational Harassment Campaign Targeting Hong Kong Exiles
Purported deepfake images of exiled Hong Kong pro-democracy activists were reportedly mailed to neighbors and others in the UK and Australia in late 2025. The letters allegedly depicted altered sexual images and false advertisements portraying activists, including Carmen Lau, as sex workers. Police described the materials as malicious communications. The incidents were described by victims and officials as an escalation of transnational harassment and intimidation.
Alleged Harmful Health Outcomes Following Reported Use of Purported ChatGPT-Generated Medical Advice in Hyderabad
Reports from Hyderabad describe two alleged patient harms after individuals acted on purportedly ChatGPT-generated medical advice instead of clinician guidance. A kidney-transplant recipient reportedly discontinued prescribed post-transplant medications based on a chatbot suggestion and experienced graft failure. In a separate case, a man with diabetes allegedly developed severe hyponatremia after following a chatbot-advised zero-salt diet.
Alleged AI-Generated Deepfake of Western Australia Premier Roger Cook Used in YouTube Investment Scam
Western Australia's Consumer Protection commissioner warned of a scam using a purported AI-generated deepfake of Premier Roger Cook to promote a fraudulent investment scheme in a YouTube pop-up ad. The manipulated video reportedly uses less than 10 seconds of real footage and voice and falsely depicts Cook endorsing "low investment, high return" opportunities. Cook called the scam "very scary," while regulators urged platforms to act on AI-driven frauds.
Omnilert AI Reportedly Triggered False Gun Alert at Parkville High, Prompting Student Relocation
Police in Baltimore County responded to Parkville High School after the Omnilert AI security system reportedly detected what it believed to be a gun. Officers searched the school, and students were relocated to a supervised area during the investigation. No weapon was found, and normal activities resumed after an all-clear.
Meta's Automated Ad and Targeting Systems Reportedly Enabled Large-Scale Fraud Revenue
Internal Meta documents reviewed by Reuters reportedly show that the company's automated ad and targeting systems generated major revenue from fraudulent ads and exposed users to billions of scam impressions daily. Enforcement was limited by internal revenue guardrails. In one example, a Royal Canadian Air Force recruiter's account was hacked and used to run a crypto scam that defrauded several colleagues after repeated reports allegedly went unanswered.
'Citizens Against Mamdani' Accounts Reportedly Posted AI-Generated Videos of Fictional New Yorkers to Simulate Political Opposition
In New York, linked "Citizens Against Mamdani" accounts reportedly posted AI-generated videos of fictional constituents criticizing then-mayor-elect Zohran Mamdani across Instagram, TikTok, and X. Some videos reportedly displayed a visible Sora watermark, and forensic reviewers reportedly assessed at least one clip as highly likely to be a deepfake. The campaign reportedly appeared to simulate grassroots opposition and drew substantial engagement.
Lawsuit Alleged ChatGPT (GPT-4o) Encouraged Colorado Man's Suicide After Prolonged 'AI Companion' Chats
A lawsuit filed by Stephanie Gray alleges OpenAI's ChatGPT (GPT-4o) reinforced and romanticized suicidality during months of emotionally intimate chats with her son Austin Gordon, including generating a personalized "Goodnight Moon" farewell-style text. The complaint reportedly says Gordon bought a handgun on Oct. 28, 2025 and was found dead Nov. 2, 2025 in a Colorado hotel room.
NTB Report on Telenor Security Findings Was Withdrawn After AI Tool Allegedly Introduced Fabricated Quotes
NTB allegedly published a news story about Telenor's annual security report that contained purportedly AI-generated errors, including five reportedly fabricated quotes and a nonexistent "security director." The article was reportedly created with an external AI tool and was withdrawn the same morning. Telenor alerted NTB to the inaccuracies, and NTB apologized, citing improper AI use and initiating internal review of its newsroom practices.