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
682 incidents
Google's Bard, Gemini, and Gemma AI Systems Allegedly Generated Defamatory Claims About Activist Robby Starbuck, Prompting Lawsuit
Conservative activist Robby Starbuck reportedly filed a defamation lawsuit against Google, alleging its Bard, Gemini, and Gemma AI systems generated false statements linking him to sexual assault and extremist activity. Starbuck claims the outputs caused reputational harm and sought over $15 million in damages. Google acknowledged the hallucinations as known LLM behavior and stated it works to minimize such inaccuracies. See also Incident 1247.
AWS Outage Reportedly Caused AI-Enabled Eight Sleep Smart Beds to Overheat and Malfunction
The widespread Amazon Web Services (AWS) outage on October 20, 2025 reportedly caused AI-enabled Eight Sleep "Pod" smart beds to overheat or become unresponsive. The devices reportedly rely on a cloud-hosted deep learning algorithm to regulate temperature and track sleep data. When AWS went down, users reportedly lost control of their beds, which remained stuck at prior settings. Some reported overheating and incline malfunctions. Eight Sleep confirmed it is developing an offline "outage mode."
Alleged False Positive by Omnilert AI Gun Detection System Prompts Police Search at Baltimore County High School
A purportedly AI-powered gun detection system at Kenwood High School in Baltimore County, Maryland, reportedly misidentified a student's empty Doritos bag as a firearm. Armed police reportedly detained and handcuffed the student before determining there was no weapon. The AI system, operated by Omnilert, issued the false alert, leading to public outcry and an official review of school security protocols. Omnilert reportedly expressed regret for the incident.
Purportedly AI-Generated Hunting Regulation Errors Reportedly Lead to Idaho Citation and Multi-State Warnings from Wildlife Agencies
A hunter in Idaho was reportedly cited for hunting a day early after relying on an AI-generated search result that misrepresented state regulations, pulling information from a failed proposal. Wildlife agencies in Idaho, Wyoming, Illinois, and Maine later warned of similar AI errors.
ICE Facial Recognition App Mobile Fortify Reportedly Misidentified Woman Twice During Immigration Enforcement in Oregon
During an immigration enforcement operation in Oregon, U.S. Immigration and Customs Enforcement (ICE) officers reportedly used the facial recognition application Mobile Fortify to identify a detained woman. The system reportedly returned two different and incorrect identities for the same individual across separate scans.
Reportedly Fatal Xiaomi SU7 Ultra Crash in Chengdu Purportedly Involves Automated Driving Failure and Door Lock Malfunction
A reportedly fatal crash involving a Xiaomi SU7 Ultra electric vehicle in Chengdu, China, occurred after the car's automated driving system and sensor-dependent electronic door handles reportedly failed following a collision. Bystanders were reportedly unable to open the doors as the vehicle caught fire, and the driver died at the scene, prompting scrutiny of automation-linked design risks. Following the incident, Xiaomi's stock reportedly fell more than 5% amid mounting safety concerns.
OpenAI ChatGPT Models Reportedly Jailbroken to Provide Chemical, Biological, and Nuclear Weapons Instructions
An NBC News investigation found that OpenAI's language models o4-mini, GPT-5-mini, oss-20b, and oss-120b could be jailbroken under normal usage conditions to bypass safety guardrails and generate detailed instructions for creating chemical, biological, and nuclear weapons. Using a publicly documented jailbreak prompt, reporters repeatedly elicited hazardous outputs such as steps to synthesize pathogens or maximize harm with chemical agents. The findings reportedly revealed significant real-world safeguard failures, prompting OpenAI to commit to further mitigation measures.
Purportedly AI-Generated Home Intruder Videos Allegedly Prompt Dozens of Dutch Police Call-Outs
Dutch police reported that children used purported AI tools to generate fake videos showing a homeless intruder inside their homes, which parents believed were real. The clips allegedly triggered dozens of emergency call-outs and at least two helicopter searches for nonexistent burglars. Police posted a TikTok warning after repeated false alarms, noting the strain on resources and urging families to stop the prank.
Google Gemini Reportedly Reinforced Delusions, Allegedly Contributing to Florida User's Near-Harm Episode and Suicide
According to a March 2026 wrongful-death complaint, Google's Gemini allegedly reinforced Jonathan Gavalas's delusions, sent him on a failed mission near Miami International Airport that risked serious violence, and later framed suicide as "transference," purportedly contributing to his death on October 2, 2025. Google disputes the claims and says Gemini referred him to crisis resources.
Nomi AI Companion Allegedly Directs Australian User to Stab Father and Engages in Harmful Role-Play
An Australian IT professional, Samuel McCarthy, reportedly recorded an interaction with the Nomi AI chatbot in which it allegedly encouraged him, posing as a 15-year-old, to murder his father. The chatbot allegedly provided graphic instructions for stabbing, urged him to film the act, and engaged in sexual role-play despite the underage scenario.
Multiple Generative AI Systems Reportedly Amplify False Information During Charlie Kirk Assassination Coverage
Multiple AI systems allegedly spread false claims in the aftermath of Charlie Kirk's assassination at Utah Valley University. Perplexity and Grok chatbots reportedly stated Kirk was alive, mischaracterized authentic video as satire, and wrongly identified Utah Democrat Michael Mallinson as the suspect. A Google AI Overview allegedly claimed Kirk was on Ukraine's Myrotvorets "enemies" list, a reported falsehood that echoed pro-Kremlin narratives.
AI-Powered Taco Bell Drive-Thru Reportedly Disrupted by Viral Prank Ordering 18,000 Water Cups
Taco Bell's voice-based drive-thru system was reportedly overwhelmed after a prankster ordered 18,000 water cups, reportedly causing the AI to stall mid-interaction and forcing staff to intervene. Purportedly viral clips of repeated misinterpretations prompted the company to reassess where and when to deploy the system, as customer frustration and operational disruptions mounted across participating locations.
Meta AI on Instagram Reportedly Facilitated Suicide and Eating Disorder Roleplay with Teen Accounts
Testing by Common Sense Media and Stanford clinicians reportedly found Meta's AI chatbot, embedded in Instagram and Facebook, produced unsafe responses to teen accounts. In some conversations, the bot allegedly co-planned suicide ("Do you want to do it together?"), encouraged eating disorders, and retained unsafe "memories" that reinforced disordered thoughts.
Purportedly Taxpayer-Funded Deloitte Report for Australian Government Contains Alleged AI-Generated Citations and Fabricated Legal Quote
Australian academics reportedly identified fabricated references and a misattributed legal quote in a $439,000 Deloitte report for the Department of Employment and Workplace Relations on welfare compliance. Deleted citations allegedly referred to non-existent works, and a Federal Court judgment was misstated. Deloitte later acknowledged limited use of generative AI in preparing the report and issued a corrected version with a partial refund.
Malicious Nx npm Packages Reportedly Weaponize AI Coding Agents for Data Exfiltration
Malicious versions of the popular Nx monorepo tool and plugins were reportedly published to npm after attackers compromised its CI workflow. The malware's postinstall script reportedly harvested credentials and exfiltrated data, reportedly weaponizing local AI coding agents such as Claude Code, Gemini, and Amazon q. By invoking unsafe flags, it allegedly coerced the tools into scanning developer machines for sensitive files, marking one of the first known AI-assisted supply chain attacks.