AI Incident Database
Browse AI incidents
Every incident record (metadata only) from the weekly snapshot of 2026-09-07. Filter, then export the selection with its licence attached; each row links to the full record and its reports.
Filter incidents (1 active)
-
Malicious OpenClaw Skills Reportedly Delivered AMOS Stealer and Exfiltrated Credentials via ClawHub AIID ↗
Bitdefender researchers reported abuse in OpenClaw's third-party 'skills' ecosystem. In a Feb. 2026 sample, about 17% of skills were reportedly assessed as malicious, with many seemingly cloned under slight name changes. Posing as utilities, some skills were reportedly found to run obfuscated commands, fetch remote payloads, and in some cases deliver AMOS Stealer on macOS. Other skills were reportedly observed searching for private keys or API tokens and exfiltrating them.
-
Moltbook Database Exposure Allegedly Revealed Users' Private Communications and API Authentication Tokens AIID ↗
Wiz researchers reported accessing an exposed Moltbook database in under three minutes, allegedly obtaining ~35,000 email addresses, thousands of private DMs, and ~1.5 million API authentication tokens. The exposure was described as enabling read/write access and potential impersonation or manipulation of "AI agent" accounts. Wiz said it disclosed the issue to Moltbook, which reportedly secured the database within hours and deleted accessed data.
-
Suspects Alleged to Have Used Purportedly AI-Generated Videos to Defraud 33 People of Approximately 15 Million Lira in Turkey AIID ↗
Turkish authorities reportedly arrested nine suspects accused of using purportedly AI-generated videos to defraud approximately 33 people of 15 million Turkish lira. Police conducted coordinated operations centered in Yozgat and extending to Antalya, Hatay, and Ankara. Six suspects were remanded in custody and three were released under judicial supervision. The specific AI tools used were not identified.
-
Grok Reportedly Generated Images After X Users Asked It to Unredact Epstein-File Photos of Children and Women AIID ↗
Bellingcat reported that, after the U.S. Department of Justice released the Epstein files, X users asked Grok to "unblur" or identify redacted children and women in sensitive photos. In 31 requests reviewed from January 30 to February 5, 2026, Grok allegedly generated images in 27 cases, despite some responses noting privacy protections. Later requests are reported to have appeared blocked after Bellingcat contacted X.
-
South Korean Woman Allegedly Used ChatGPT to Assess Lethality of Drug-and-Alcohol Mixtures Before Two Fatal Motel Poisonings AIID ↗
In Seoul, a woman allegedly used ChatGPT to ask whether mixing sleeping pills or benzodiazepines with alcohol could be fatal before poisoning drinks given to three men. Two men later died in separate motel incidents, and a third survived after losing consciousness. Police reportedly cited her chatbot queries and search history as evidence of intent.
-
Purportedly AI-Altered Images Reportedly Distort Evidence After Minneapolis Shooting of ICU Nurse Alex Pretti AIID ↗
Following the fatal shooting of Minneapolis ICU nurse Alex Pretti by U.S. Customs and Border Patrol agents, social media accounts reportedly circulated images purported to have been altered by AI, reportedly distorting evidence of the incident by portraying Pretti as threatening law enforcement and altering the presence of weapons. The images reportedly misidentified individuals and helped reinforce partisan narratives by purportedly obscuring verified video and eyewitness accounts.
-
NPR Host David Greene Alleged Google's NotebookLM Replicated His Voice Without Consent, Prompting Lawsuit AIID ↗
NPR's David Greene reportedly sued Google LLC and Alphabet in Santa Clara County, alleging NotebookLM's Audio Overviews uses a synthetic male voice that purportedly mimics his cadence and delivery without consent or compensation. The complaint reportedly cited an independent voice-recognition analysis reporting 53–60% confidence his voice trained the model. Google reportedly called the allegations baseless and said the voice is based on a paid actor.
-
Waymo Autonomous Vehicle Reportedly Struck Child Near Elementary School in Santa Monica, California AIID ↗
A Waymo driverless vehicle reportedly struck a child near an elementary school in Santa Monica, California during school drop-off hours. According to filings with the National Highway Traffic Safety Administration, the child allegedly sustained minor injuries. The vehicle was purportedly operating on Waymo's 5th Generation Automated Driving System without a human safety driver. Waymo reported the incident to federal regulators, who subsequently opened an investigation.
-
White House Reportedly Shares Purportedly AI-Altered Arrest Photo Depicting Minnesota Protester Nekima Levy Armstrong as Crying AIID ↗
The White House reportedly posted a purportedly AI-altered image on X showing Minnesota protester and attorney Nekima Levy Armstrong appearing to cry during her arrest. An earlier image reportedly shared by Homeland Security Secretary Kristi Noem showed her calm. Reported analysis by third parties using AI detection tools found signs of purported facial manipulation, which could be replicated using generative AI systems.
-
Purportedly AI-Generated Tasmania Tours Content Reportedly Misled Tourists Into Traveling to Nonexistent Weldborough Hot Springs AIID ↗
A purportedly AI-generated travel article on the Tasmania Tours website reportedly invented "Weldborough Hot Springs" and promoted them as a real attraction. Tourists allegedly traveled to the remote town of Weldborough in search of the site, while local businesses reportedly received repeated calls and visits from confused travelers asking how to find the nonexistent springs.
-
DHS Agents Reportedly Threatened Legal Observers With 'Domestic Terrorist' Database While Using Purportedly AI-Enabled Surveillance During ICE Operations AIID ↗
Two plaintiffs alleged that DHS personnel used purportedly AI-enabled surveillance to identify people recording immigration enforcement and threatened to add them to a "domestic terrorist" database or watch list. The complaint reportedly includes video in which an agent appears to reference a "database," and another alleged warning that agents would "come to your house later tonight." DHS has reportedly denied that such a database exists while acknowledging threat monitoring.
-
Eightfold AI Hiring Tools Allegedly Secretly Scored Job Applicants for Employers AIID ↗
Two California job applicants reportedly filed a proposed class action alleging that Eightfold AI's hiring platform secretly gathered personal data, generated candidate reports, and scored applicants for employers without required notice or dispute rights. The complaint says Eightfold's Evaluation Tools ranked applications submitted to companies including PayPal and Microsoft; both named plaintiffs alleged they were evaluated and rejected or not advanced.
-
Scammers Reportedly Used AI-Generated Image of Missing Puppy Hazel to Solicit Fraudulent Vet Payment from St. Petersburg, Florida Couple AIID ↗
Scammers reportedly used an AI-generated or AI-altered image of a missing German Shepherd puppy on an operating table to convince a St. Petersburg, Florida couple that the dog had been hit by a car and needed emergency surgery. The caller allegedly impersonated police and veterinary staff while requesting $1,900 through Zelle. The payment was flagged as fraud, and the dog later returned home unharmed.
-
Purportedly AI-Generated Memorial Posters Reportedly Presented Fabricated Child Portraits as Victims of South African School Transport Crash AIID ↗
Purportedly AI-generated memorial posters reportedly presented fabricated child portraits as pupils killed in a school-transport crash near Vanderbijlpark, South Africa. The posters circulated on Facebook, X and TikTok before authorities released the victims' identities or photographs. AFP reported that the depicted faces did not match official victim images, and the Gauteng education department condemned the posts while families were grieving.
-
Spokane Transit Authority Onboard Navigation System Reportedly Routed Double-Decker Bus to Low Bridge, Injuring Seven AIID ↗
In Spokane, Washington, a Spokane Transit Authority double-decker bus was reportedly routed by its onboard navigation/CAD mapping system onto Cedar Street and hit a low-clearance railroad viaduct, shearing the upper deck. Of 10 people aboard, seven were reportedly hospitalized with non-life-threatening injuries. STA reportedly later warned operators not to use CAD maps for routing and pulled double-deckers from service pending an investigation; the driver was cited for negligent driving.
-
Zoox Robotaxi Reportedly Collided with Open Car Door, Allegedly Injuring San Francisco Street Ambassador AIID ↗
A Zoox robotaxi reportedly collided with the open driver-side door of a parked car in San Francisco on January 17, 2026, allegedly injuring street ambassador Jamel Durden's hand. Zoox reportedly said the door was opened suddenly and that the vehicle detected it and attempted to avoid contact. Police opened an investigation, and the California DMV reviewed the crash.
-
Lower Saxony CDU Employee Allegedly Shared Sexualized Purported Deepfake of Colleague in Internal WhatsApp Group AIID ↗
A senior employee of the CDU parliamentary faction in Lower Saxony allegedly posted a purportedly AI-generated sexualized deepfake of a female colleague in an internal WhatsApp group. Prosecutors later said the video appeared to be an AI montage created by inserting a real image of the woman into a bikini-dance clip. The employee was reportedly later dismissed.
-
AI incident: sexual content (Jan 2026) AIID ↗
An AI Incident Database record involving sexual content, dated 17 January 2026. Risk classification, related incidents and the laws on record for this risk area.
-
Coco Robotics Delivery Robot Reportedly Became Stuck on Railroad Tracks and Was Struck by Train in Miami AIID ↗
A Coco Robotics delivery robot reportedly became stranded on railroad tracks in Miami and was struck and destroyed by a passing train after remaining stalled for about 15 minutes. Local and secondary reports said no one was injured. Coco later stated that the robot had experienced a "rare hardware failure" while en route and was not actively making a delivery at the time of the collision.
-
Purported AI-Generated Images Depict Kate Garraway With Fictitious Partner AIID ↗
ITV presenter Kate Garraway reported that purported AI-generated images falsely depicting her in a relationship with a fictitious partner circulated online, prompting false rumors about her personal life. Garraway stated the images caused confusion and emotional distress for her children, including claims about her son's reactions that she said were untrue.
-
ICE AI Resume Screening Error Allegedly Routed Inexperienced Recruits Into Inadequate Training Pathways AIID ↗
U.S. Immigration and Customs Enforcement (ICE) reportedly used an AI-assisted résumé screening tool during a 2025 hiring surge that misclassified some applicants as having law-enforcement experience. As a result, certain recruits without policing backgrounds were allegedly routed into a shortened training pathway. ICE reportedly identified the error, reviewed résumés manually, and reassigned affected recruits for additional training.
-
Automated Shuttle Bus Was Reportedly Rear-Ended During U.S. Department of Transportation Demonstration Ride in Washington, D.C. AIID ↗
An automated shuttle bus operated by Beep was reportedly involved in a minor collision during a U.S. Department of Transportation demonstration ride in Washington, D.C. The vehicle, operating autonomously with a human safety driver onboard, was reportedly rear-ended by a Tesla whose driver made an illegal lane change. No injuries were reported, and officials stated the autonomous system functioned appropriately.
-
Border Patrol Agent Allegedly Claimed Facial Recognition Identified Minneapolis ICE Observer and Global Entry Was Reportedly Revoked Three Days Later AIID ↗
On 01/10/2026 near Minneapolis, Minnesota, legal observer Nicole Cleland reportedly stated that a CBP Border Patrol agent stopped her vehicle, addressed her by name, and claimed agents used facial recognition to identify her while recording on body cam. She reportedly had her Global Entry and TSA PreCheck revoked on 01/13/2026, which she alleged was retaliatory intimidation, causing travel burden and chilling effects.
-
Purported AI-Generated Doctor Deepfakes Reportedly Used Guy's and St Thomas' Branding to Market Weight Loss Patches AIID ↗
Guy's and St Thomas' NHS Foundation Trust warned that purported AI-generated videos circulated on Facebook and TikTok depicted its clinicians endorsing weight loss patches. The videos allegedly impersonated doctors and used misleading medical claims to market a product.
-
Purported AI Voice Clone Allegedly Narrated Shaun Rein's 'The Split' in Unauthorized YouTube 'Podcast' Videos AIID ↗
Author Shaun Rein reported that seven YouTube "podcast" videos on the YouTube channel "The US-China Narrative" used an allegedly AI-cloned version of his voice to narrate material taken from his 2024 book The Split, without his or his publisher's permission. The videos reportedly drew 200,000+ streams in a week.
Disclaimer: informational only, not legal advice. Verify every claim against the linked official sources and consult a qualified lawyer before acting.
Frequently asked questions
What is recorded in each incident?
What happened, when, who was involved as developer or deployer, who was harmed, the sectors and countries affected, and the domain and subdomain the harm falls under. Incidents are mirrored from the AI Incident Database and each links back to its source record.
Does an incident mean wrongdoing was proven?
No. These are reported incidents, not findings of liability. Entities named in a report are named as reported, allegations are marked as alleged, and nothing here is a legal determination.
How current is the incident data?
It syncs directly from the AI Incident Database several times a day, so a new incident usually appears within hours. Each record carries the date it was last synced, and the snapshot date is shown on the page.
Does this page include the full incident reports?
No. The report texts are outside the licence this data is shared under, so only the metadata and the classifications appear here. Every row links to the original incident record, which carries the reports themselves.