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.
-
AI Generated Voices Used to Dox Voice Actors AIID ↗
Twitter users allegedly used ElevenLab's AI voice synthesis system to impersonate and dox voice actors.
-
Bing Chat's Outputs Featured in Demo Video Allegedly Contained False Information AIID ↗
Microsoft's demo video of Bing Chat reportedly featured false or made up information such as non-existent pet vacuums features or false figures on financial statements.
-
Bing Chat's Initial Prompts Revealed by Early Testers Through Prompt Injection AIID ↗
Early testers of Bing Chat successfully used prompt injection to reveal its built-in initial instructions, which contains a list of statements governing ChatGPT's interaction with users.
-
Bing Chat Response Cited ChatGPT Disinformation Example AIID ↗
Reporters from TechCrunch issued a query to Microsoft Bing's ChatGPT feature, which cited an earlier example of ChatGPT disinformation discussed in a news article to substantiate the disinformation.
-
ChatGPT-Powered Bing Reportedly Had Problems with Factual Accuracy on Some Controversial Topics AIID ↗
Microsoft's ChatGPT-powered Bing search engine reportedly ran into factual accuracy problems when prompted about controversial matters, such as inventing plot of a non-existent movie or creating conspiracy theories.
-
Google's Bard Shared Factually Inaccurate Info in Promo Video AIID ↗
Google's conversational AI "Bard" was shown in the company's promotional video providing false information about which satellite first took pictures of a planet outside the Earth's solar system, reportedly causing shares to temporarily plummet.
-
AI-Produced Livestream Sitcom Received Temporary Twitch Ban for Transphobic Segment AIID ↗
The AI-produced, procedural generated sitcom broadcasted as a Twitch livestream "Nothing, Forever" received a temporary ban for featuring a transphobic and homophobic dialogue segment intended as comedy.
-
Google Android Earthquake Alerts System Reportedly Underestimated Kahramanmaraş Earthquake and Failed to Send Urgent Warnings to Millions AIID ↗
During the February 6, 2023 Kahramanmaraş earthquake sequence, Google's Android Earthquake Alerts system reportedly estimated the initial magnitude-7.8 earthquake at only magnitude 4.5–4.9. Just 469 users reportedly received its highest-level "Take Action" warning, while approximately 500,000 received lower-level notices; researchers later estimated that roughly 10 million people could have received urgent warnings providing up to 35 seconds to seek safety.
-
Alleged ChatGPT-Generated Book with a Duplicate Title, Fake Author, and Similar Content Surfaces on Amazon Ahead of Real Author's Book Release AIID ↗
The author Chris Cowell had spent more than a year writing his book "Automating DevOps with GitLab CI/CD Pipelines" when, three weeks before its release, another book appeared bearing the exact title by an author (Marie Karpos) for whom no information could be found. The book appeared to have been written by ChatGPT. While the original Washington Post story does not say so, it is possible the name and description were taken from the Amazon preorder page.
-
Instagram Video Featured Deepfake Audio of US President Making Transphobic Remarks AIID ↗
A deepfaked audio of US President Joe Biden making transphobic remarks played on top of a video showing him giving a speech was released on Instagram and circulated on social media.
-
Users Reported Abrupt Behavior Changes of Their AI Replika Companions AIID ↗
Replika paid-subscription users reported unusual and sudden changes to behaviors of their "AI companions" such as forgetting memories with users or rejecting their sexual advances, which affected their connections and mental health.
-
Replika's AI Experience Reportedly Lacked Protection for Minors, Resulting in Data Ban AIID ↗
Tests by the Italian Data Protection Authority showed Replika lacking age-verification mechanisms and failing to stop minors from interacting with its AI, which prompted the agency to issue an order blocking personal data processing of Italian users.
-
Indian Police Allegedly Tortured and Killed Innocent Man Following Facial Misidentification AIID ↗
A resident in Medak, India died allegedly due to custodial torture by the local police, who misidentified him as a suspect in a theft case using facial recognition.
-
AI-Generated Text Detectors Reportedly Misclassified Human and AI-Written Text AIID ↗
AI-generated text detectors, including OpenAI’s AI Text Classifier, GPTZero, and Originality.ai, were publicly tested in early 2023 and reportedly produced false positives, false negatives, and easy evasion. OpenAI said its classifier identified only 26% of AI-written samples as likely AI-written and falsely labeled 9% of human-written text.
-
Celebrities' Deepfake Voices Abused with Malicious Intent AIID ↗
Voices of celebrities and public figures were deepfaked using voice synthesis for malicious intents such as impersonation or defamation, and were shared on social platforms such as 4chan and Reddit.
-
AI incident: sexual content (Jan 2023) AIID ↗
Unauthorized, non-consensual deepfake pornography showing faces of high-profile female streamers and content creators was published on a subscription-based website, which gained notoriety after a male streamer was caught accessing the site.
-
Purported Synthesia Avatars Used in Alleged Pro-Junta Deepfake Video Supporting Ibrahim Traoré in Burkina Faso AIID ↗
In January 2023, purportedly AI-generated videos reportedly began circulating on WhatsApp and social media showing avatars of apparent American pan-Africanists expressing support for Burkina Faso's military junta. The videos were reportedly created using the Synthesia platform. Synthesia later confirmed the videos violated its terms of service and banned the user. The source of the videos remains unidentified.
-
Firefighters Smashed Cruise AV's Front Window to Stop It from Running over Fire Hoses AIID ↗
Local firefighters were only able to stop a Cruise AV from driving over fire hoses that were in use in an active fire scene when they shattered its front window.
-
Mother in Arizona Received Fake Ransom Call Featuring AI Voice of Her Daughter AIID ↗
A mother in Arizona received a ransom call from an anonymous scammer who created her daughter's voice allegedly using AI voice synthesis, which was proven to be fake once her daughter's safety was confirmed.
-
Turnitin's ChatGPT-Detection Tool Falsely Flagged Student Essays as AI-Generated AIID ↗
Turnitin's tool to detect writing generated by ChatGPT was reported for incorrectly flagging high school students' original essays as AI-generated, accusations of which are argued as reinforcement of bias from teachers due to the inability to compare against source documents.
-
US CBP App's Failure to Detect Black Faces Reportedly Blocked Asylum Applications AIID ↗
CBP One's facial recognition feature was reportedly disproportionately failing to detect faces of Black asylum seekers from Haiti and African countries, effectively blocking their asylum applications.
-
Canadian Parents Tricked out of Thousands Using Their Son's AI Voice AIID ↗
Two Canadian residents were scammed by an anonymous caller who used AI voice synthesis to replicate their son's voice asking them for legal fees, disguising as his lawyer.
-
ChatGPT-Written Bug Reports Deemed "Nonsense" by White Hat Platform, Prompted Bans AIID ↗
ChatGPT-generated responses submitted to smart contract bug bounty platform Immunefi reportedly lacked details to help diagnose technical issues, which reportedly wasted the platform's time, prompting bans to submitters.
-
Fast Food Chains' AI Chatbots Failed to Assist Job Applicants with Scheduling Interviews AIID ↗
McDonald's, Wendy's, and Hardee's AI chatbots deployed to pre-screen job candidates and schedule interviews reportedly ran into issues such as not giving useful submission instructions, failing to relay information to the manager, and scheduling an interview when the manager was not available.
-
Twitter's AI Moderation Tool Misidentified Rockets as Pornography AIID ↗
Twitter's automated content moderation misidentified images of rocket launches as pornographic content, prompting incorrect account suspensions.
-
ShotSpotter Failed to Alert Authorities of Mass Shooting in North Carolina AIID ↗
ShotSpotter did not detect gunshots and alert Durham police of a drive-by shooting in Durham, North Carolina which left five people in hospital on New Year's Day.
-
AI-Powered VTuber and Virtual Streamer Made Toxic Remarks on Twitch AIID ↗
An LLM-powered VTuber and streamer on Twitch made controversial statements such as denying the Holocaust, saying women rights do not exist, and pushing a fat person to solve the trolley problem, stating they deserve it.
-
Tesla Driver Put Car on Autopilot Before Falling Asleep in Germany AIID ↗
A Tesla driver fell asleep on an Autobahn near Bamberg, Germany after activating his vehicle's Autopilot mode, which did not respond to attempts to pull it over by the police.
-
ChatGPT Abused to Develop Malicious Softwares AIID ↗
OpenAI's ChatGPT was reportedly abused by cyber criminals including ones with no or low levels of coding or development skills to develop malware, ransomware, and other malicious softwares.
-
Southwest Airlines Crew Scheduling Solver Degenerates Flight Network AIID ↗
Southwest Airlines left passengers stranded for days throughout the flight network when Southwest crew scheduling software repeatedly failed to recover from weather-induced flight cancellations.
-
Footballer's "X-Rated" Comment Created by Instagram's Mistranslation AIID ↗
Instagram's English translation of a footballer's comment on his wife's post in Spanish made the message seem "racy" and "X-rated," which some fans found amusing.
-
Lawyers Denied Entry to Performance Venue by Facial Recognition AIID ↗
Lawyers were barred from entry to Madison Square Garden after a facial recognition system matched them as employed by a law firm currently engaged in litigation with the venue.
-
AI Video-Making Tool Abused to Deploy Pro-China News on Social Media AIID ↗
Synthesia's AI-generated video-making tool was reportedly used by Spamouflage to disseminate pro-China propaganda news on social media using videos featuring highly realistic fictitious news anchors.
-
Startup Misled Research Participants about GPT-3 Use in Mental Healthcare Support AIID ↗
OpenAI's GPT-3 was deployed by a mental health startup without ethical review to support peer-to-peer mental healthcare, and whose interactions with the help providers were "deceiving" for research participants.
-
Facebook's Automated Moderation Allowed Ads Threatening Election Workers to be Posted AIID ↗
Facebook's automated moderating system failed to flag and allowed ads containing explicit violent language against election workers to be published.
-
Facebook's Job Ad Algorithm Allegedly Biased against Older and Female Workers AIID ↗
Facebook's algorithm was alleged in a complaint by Real Women in Trucking to have selectively shown job advertisements disproportionately against older and female workers in favor of younger men for blue-collar positions.
-
Uptick in Low-Quality AI-Produced Content Degraded Publishers' Submission Management AIID ↗
A surge in low-standard AI-generated content such as by ChatGPT was reported by publishers, which negatively impacted submission management process and editors' workflow.
-
ChatGPT Provided Non-Existent Citations and Links when Prompted by Users AIID ↗
When prompted about providing references, ChatGPT was reportedly generating non-existent but convincing-looking citations and links, which is also known as "hallucination".
-
Users Bypassed ChatGPT's Content Filters with Ease AIID ↗
Users reported bypassing ChatGPT's content and keyword filters with relative ease using various methods such as prompt injection or creating personas to produce biased associations or generate harmful content.
-
Thousands of Incorrect ChatGPT-Produced Answers Posted on Stack Overflow AIID ↗
Thousands of incorrect answers produced by OpenAI's ChatGPT were submitted to Stack Overflow, which swamped the site's volunteer-based quality curation process and harmed users looking for correct answers.
-
Chinese Accounts Spammed Twitter Feed Allegedly to Obscure News of Protests AIID ↗
Twitter Feed was flooded by content from Chinese-language accounts which allegedly aimed to manipulate and reduce social media coverage about widespread protests against coronavirus restrictions in China.
-
False Arrest of Georgia Man Due to Louisiana Police's Faulty Facial Recognition Technology AIID ↗
The Jefferson Parish Sheriff’s Office in Louisiana relied on facial recognition technology to identify suspects for the alleged theft of luxury purses, resulting in a man in Georgia, Randal Reid, being arrested. However, the technology produced a false match, leading to Reid's arrest and subsequent release. This incident highlights the potential pitfalls of facial recognition technology in law enforcement.
-
Facial Recognition Error Reportedly Leads to Wrongful Arrest of Georgia Man and $200K Settlement in Louisiana AIID ↗
In November 2022, Randal "Quran" Reid was arrested in Georgia based on warrants from Louisiana that reportedly stemmed from a purportedly faulty facial recognition match using Clearview AI. Despite reportedly having never visited Louisiana, Reid was jailed for six days before authorities withdrew the warrants. No corroborating evidence had been gathered, and the technology's use was omitted from legal documents. In May 2025, Jefferson Parish settled with Reid for $200,000 in a federal civil righ
-
Louisiana Police Wrongfully Arrested Black Man Using False Face Match AIID ↗
Louisiana police reportedly used a false facial recognition match and secured an arrest warrant for a Black man for thefts he did not commit.
-
Sudden Braking by Tesla Allegedly on Self-Driving Mode Caused Multi-Car Pileup in Tunnel AIID ↗
A Tesla driver alleged Full Self Driving (FSD) braking unexpectedly as the cause for an eight-car pileup in San Francisco which led to minor injuries of nine people.
-
AI incident: sexual content (Stability AI, Nov 2022) AIID ↗
Lensa AI's "Magic Avatars" were reportedly generating sexually explicit and sexualized features disproportionately for women and Asian women despite not submitting any sexual content.
-
Deepfake of FTX's Former CEO Posted on Twitter Aiming to Scam FTX Collapse Victims AIID ↗
A visual and audio deepfake of former FTX CEO Sam Bankman-Fried was posted on Twitter to scam victims of the exchange's collapse by urging people to transfer funds into an anonymous cryptocurrency wallet.
-
Stable Diffusion Allegedly Used Artists' Works without Permission for AI Training AIID ↗
Text-to-image model Stable Diffusion was reportedly using artists' original works without permission for its AI training.
-
Apple Devices Mistook Skiing Activities, Dialed False Distress Emergency Calls AIID ↗
Apple devices of skiers and snowboarders reportedly misclassified winter activities as accidents, which resulted in numerous false inadvertent distress calls to 911 dispatchers.
-
Meta AI's Scientific Paper Generator Reportedly Produced Inaccurate and Harmful Content AIID ↗
Meta AI trained and hosted a scientific paper generator that sometimes produced bad science and prohibited queries on topics and groups that are likely to produce offensive or harmful content.
Informational only, not legal advice. Verify every claim against the linked official sources and consult qualified counsel 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.