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
AI-Produced Livestream Sitcom Received Temporary Twitch Ban for Transphobic Segment
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.
Instagram Video Featured Deepfake Audio of US President Making Transphobic Remarks
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.
Alleged ChatGPT-Generated Book with a Duplicate Title, Fake Author, and Similar Content Surfaces on Amazon Ahead of Real Author's Book Release
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.
Indian Police Allegedly Tortured and Killed Innocent Man Following Facial Misidentification
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.
Replika's AI Experience Reportedly Lacked Protection for Minors, Resulting in Data Ban
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.
AI-Generated Text Detectors Reportedly Misclassified Human and AI-Written Text
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.
Non-consensual Deepfake Porn Targeted Female Content Creators
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.
Firefighters Smashed Cruise AV's Front Window to Stop It from Running over Fire Hoses
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.
Turnitin's ChatGPT-Detection Tool Falsely Flagged Student Essays as AI-Generated
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.
Mother in Arizona Received Fake Ransom Call Featuring AI Voice of Her Daughter
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.
US CBP App's Failure to Detect Black Faces Reportedly Blocked Asylum Applications
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.
Fast Food Chains' AI Chatbots Failed to Assist Job Applicants with Scheduling Interviews
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
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
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
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.