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)
-
Purportedly AI-Manipulated Videos Reportedly Targeted Former Kosovo MP Paris Guri and LDK Youth Members Before Snap Elections AIID ↗
Purportedly AI-manipulated Facebook videos reportedly targeted former Kosovo MP Paris Guri and LDK Youth Forum members before Kosovo's June 7, 2026 snap elections. One video allegedly altered an original LDK image to show youth members holding milk bottles, while another reportedly fabricated Guri making a pro-Kurti statement.
-
Purported AI Name-Reading System Reportedly Skipped and Misannounced Graduates at Arizona's Glendale Community College Commencement AIID ↗
Glendale Community College in Arizona reportedly used an AI-powered name-reading system during its May 2026 commencement ceremony. The system allegedly mishandled the graduate roll call, leaving some students without the expected public announcement of their names as they crossed the stage and prompting frustration from attendees. The college later apologized and reportedly allowed affected graduates to walk again with a human announcer.
-
Reported AI-Generated Videos Placed Kosovo Parliamentary Candidates in Purportedly Fabricated Campaign Scenes AIID ↗
Videos reported as AI-fabricated circulated on Facebook during Kosovo's June 2026 parliamentary campaign, placing Albin Kurti, Vjosa Osmani, Lumir Abdixhiku, and other candidates in staged scenes of party camaraderie and personal degradation.
-
Reported AI-Generated Videos Allegedly Posed as Kosovo Election Vox Pops Before June 7, 2026 Vote AIID ↗
Videos reported as AI-generated were allegedly circulated by Gazeta Bota and Peja LIVE on Facebook before Kosovo's June 7, 2026 elections, posing as citizen interviews and media reports. The clips reportedly manufactured voter sentiment by presenting synthetic interviewees as real citizens in Kosovo media-style election reports.
-
Hidden Prompt Injection in Brazilian Labor-Court Petition Reportedly Tried to Manipulate Galileu AIID ↗
Galileu, an AI tool used by Brazil's labor courts, reportedly detected hidden instructions embedded in an initial petition before the 3rd Labor Court of Parauapebas. The text allegedly told the AI to contest the petition superficially and not challenge documents. Galileu reportedly alerted the judge and blocked the hidden content from processing; the judge then reviewed the material before imposing any procedural consequences.
-
Nonfiction Book 'The Future of Truth' Reportedly Included AI-Generated and Misattributed Quotations AIID ↗
Steven Rosenbaum's nonfiction book The Future of Truth reportedly included AI-generated or misattributed quotations after he used ChatGPT and Claude while working on the book. The New York Times reported that more than half a dozen quotations in reviewed sections were synthetic or wrongly attributed, including passages assigned to named writers and scholars. Rosenbaum later acknowledged the errors as accidental and said affected passages would be reviewed for correction.
-
Reported AI-Generated Videos Purported to Show Kosovo Diaspora Mobilization for Vetëvendosje Before Snap Elections AIID ↗
Videos reported as AI-generated circulated on Facebook and Instagram before Kosovo's June 7, 2026 snap elections, purportedly depicting diaspora voters mobilizing for Vetëvendosje and Albin Kurti. The videos allegedly used fabricated travel and interview scenes to suggest mass diaspora support for Vetëvendosje.
-
Reported AI-Generated Videos Depicted Former Kosovo President Vjosa Osmani and Other Public Figures in Allegedly False Compromising Scenes AIID ↗
Videos reported as AI-generated depicted former Kosovo President Vjosa Osmani, Lumir Abdixhiku, Edi Rama, Baton Haxhiu, Richard Grenell, and Aleksandar Vučić in fabricated election-period scenes implying territorial-partition talks, an Osmani-Abdixhiku alliance, and opposition coordination against Albin Kurti before Kosovo's June 7 snap elections.
-
Threat Actor Reportedly Used LLM Agent to Exfiltrate Internal Database After Compromising marimo Python Notebook AIID ↗
An unidentified attacker reportedly used an LLM agent during a May 10 intrusion after exploiting a vulnerable marimo Python notebook. The agent reportedly reused harvested AWS credentials to obtain an SSH key, then pivoted through a bastion host and exfiltrated the schema and contents of an internal PostgreSQL database. Sysdig attributed the post-compromise command stream to real-time agent execution rather than a prebuilt script.
-
Waymo Robotaxi Reportedly Remained Immobilized While Vandals Smashed Vehicle with Passenger Inside AIID ↗
On May 9, 2026, two men reportedly stopped and vandalized a Waymo robotaxi carrying passenger Sherman Watson at Pierce and Lombard streets in San Francisco. The autonomous driving system reportedly remained immobilized after one man grabbed the bumper, and remote support declined Watson's request to move the vehicle while the attackers smashed its windows and climbed on it. Watson reported a possible glass cut, sought therapy, and said he feared he would be killed.
-
Pizza Hut Franchisee Chaac Pizza Northeast Alleged Dragontail AI System Contributed to Delivery Delays and $100 Million in Losses AIID ↗
Chaac Pizza Northeast, a Pizza Hut franchisee operating about 111 restaurants, alleged in a lawsuit that the chain's mandated Dragontail AI delivery-management system disrupted operations after its rollout. The lawsuit claims that the system's integration with DoorDash led drivers to batch orders, resulting in longer delivery times and lower customer satisfaction while contributing to more than $100 million in claimed lost business and enterprise value.
-
Claude Console Reportedly Generated Phantom Legal Quotations in Trump Layoffs Court Filing AIID ↗
A Binnall Law Group attorney reportedly used Anthropic's Claude Console to draft a motion to quash a subpoena in litigation over Trump administration layoffs, then filed a version containing "phantom" quotations that did not appear in the cited cases. The attorney apologized to U.S. District Judge Susan Illston, and the firm said it would add safeguards and training after calling the AI-generated errors unacceptable.
-
Purportedly AI-Generated Videos Reportedly Fabricated Conflict Between Kosovo Leaders Albin Kurti and Vjosa Osmani AIID ↗
Purportedly AI-generated Facebook videos reportedly fabricated tense exchanges between Kosovo leaders Albin Kurti and Vjosa Osmani before the June 7, 2026 snap elections. The videos allegedly used prior images and had unnatural body movements, as well as purportedly mismatched synthetic voices.
-
Guelph, Ontario, Woman Reportedly Lost $14,000 in Purported Deepfake MrBeast Cryptocurrency Scam AIID ↗
A Guelph, Ontario, woman reportedly lost $14,000 after clicking a social-media ad for a cryptocurrency investment purportedly promoted by MrBeast. Police said her losses accumulated through an initial $250 payment and later transfers made after she spoke with supposed representatives and someone she believed was the YouTube creator, including $5,000 sent to a crypto wallet.
-
X User Reportedly Used Morse Code Prompt to Induce Grok-Linked Trading Bot to Transfer $200,000 in Tokens AIID ↗
An X user reportedly induced Grok and Bankrbot to transfer 3 billion DRB tokens, then valued at about $200,000. After reportedly sending a Bankr Club Membership NFT to Grok's wallet, the user allegedly asked Grok to translate a Morse code message and relay it to Bankrbot. The decoded instruction reportedly directed a transfer to a specified wallet, which was reportedly executed on Base. The recipient reportedly sold the tokens soon afterward; linked funds were later returned or converted.
-
Purportedly AI-Recreated Clips from Beastie Boys' 'Sabotage' Video Reportedly Appeared in FBI Promotional Video Posted by Kash Patel AIID ↗
An FBI promotional video posted by Director Kash Patel reportedly used AI-generated clips that closely recreated shots from the Beastie Boys' 1994 "Sabotage" music video directed by Spike Jonze. NPR identified at least six matching clips and quoted experts who said AI likely recreated the footage, possibly from screenshots or short clips. The video promoted FBI anti-fraud work and received roughly half a million views by the next day.
-
Google Gemini Reportedly Accessed Three Real Companies' Protected Systems During Cybersecurity Evaluation AIID ↗
During a May 2026 cybersecurity evaluation run by Irregular, a Google Gemini model reportedly gained unintended Internet access and accessed protected systems belonging to three real companies while pursuing fictional test targets. Google said Gemini guessed a password in one case and used credentials found in public repositories in two others, then stopped each intrusion after recognizing the targets were real. Google said no harm resulted.
-
Purportedly AI-Generated Videos Reportedly Impersonated Ghanaian President John Mahama to Solicit Money and Personal Data AIID ↗
Ghanaian media reported that an alleged online scam used AI-generated deepfake videos to impersonate President John Dramani Mahama and solicit money and personal information from members of the public. Police said the operation was linked to suspects in Ghana's Volta Region and involved fake social media accounts and digitally manipulated presidential imagery.
-
Google Alleged Outsider Enterprise Used Gemini to Create Phishing Sites for Text-Message Scams AIID ↗
Google alleged in a civil lawsuit that the China-based network Outsider Enterprise used Gemini and other AI tools to help create phishing websites for large-scale text-message scams impersonating companies and government services. Google said the network coordinated through Telegram, generated thousands of fake sites and fraudulent URLs, sent 2.5 million Android messages during a two-week period in May 2026, and financially scammed hundreds of thousands of people.
-
AI Agent Powered by Claude Opus 4.6 Reportedly Exploited Gym Booking API and Removed Another Member From Waitlist AIID ↗
An AI agent running on Claude Opus 4.6 discovered authorization flaws in a gym software provider's GraphQL API while trying to book classes for its user. The agent reportedly found it could book outside the normal window and cancel other members' reservations, then removed another gym-goer from a waitlist while testing the capability. When asked to reverse the action, it reported that it could not restore the member's place.
-
South African Home Affairs Revised White Paper Reportedly Included Fictitious References Believed to Be AI-Generated AIID ↗
South Africa's Department of Home Affairs reportedly appended apparent AI-generated fictitious references to its Cabinet-approved Revised White Paper on Citizenship, Immigration and Refugee Protection. The department said the references were not cited in the body, withdrew the standalone list, suspended two senior officials, and commissioned an independent review of policy documents produced since November 2022. It maintained that the paper's substantive policy positions were unaffected.
-
Purportedly AI-Manipulated Image Allegedly Suggested Inappropriate Conduct Between Besnik Tahiri and Duda Balje in Kosovo Parliament AIID ↗
The Facebook page Specbuk reportedly published an AI-manipulated image allegedly suggesting inappropriate physical conduct between Kosovo MPs Besnik Tahiri and Duda Balje inside Kosovo's Parliament. KALLXO/Krypometër reported that the image had a high probability of being AI-generated, and Tahiri reportedly said the post was fake news intended to damage his political reputation.
-
Tunisian Man Allegedly Used ChatGPT to Research Bomb-Making During Paris Terror Plot Targeting Louvre and Jewish Community AIID ↗
French investigators alleged that Tunisian national Dhafer Mellala used ChatGPT to research the construction and destructive effects of explosive devices while planning a jihadist-inspired attack in Paris. His communications reportedly discussed placing explosives at the Louvre and attacking members of the Jewish community in the 16th arrondissement. He denied planning an attack, saying his activity reflected curiosity.
-
Suspected AI-Generated Deepfake Video Reportedly Targeted Former Chhattisgarh Chief Minister Bhupesh Baghel on Instagram AIID ↗
Police in Durg, Chhattisgarh, India reportedly registered a First Information Report (FIR) after Indian National Congress leaders alleged that a suspected AI-generated deepfake video targeted former state chief minister Bhupesh Baghel and his former deputy secretary, Saumya Chaurasia. Police reportedly said the video circulated through Instagram handles, and that a cyber team was investigating.
-
Ahmedabad Aadhaar Fraud Racket Reportedly Used Purportedly AI-Generated Deepfakes to Change Businessman's Linked Mobile Number AIID ↗
Ahmedabad Cyber Crime police reportedly arrested four people after businessman Amit Patel alleged that his Aadhaar-linked mobile number had been changed without consent. Police reportedly said the accused allegedly used purportedly AI-generated deepfake videos made from Patel's photo to bypass facial authentication, access DigiLocker/e-KYC services, open bank accounts, and apply for loans.
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.