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
1,751 incidents
Auto-Insurance Photo-Based Estimation Allegedly Gave Inaccurate Repair Prices Frequently
Auto-insurance companies' photo-based estimation of repair price was alleged by repair shop owners and industry groups as providing inaccurate estimates, causing damaged cars to stay in the shop longer.
Black Uber Eats Driver Allegedly Subjected to Excessive Photo Checks and Dismissed via FRT Results
A lawsuit by a former Uber Eats delivery driver alleged the company to have wrongfully dismissed him due to frequent false mismatches of his verification selfies, and discriminated against him via excessive verification checks.
Players Manipulated GPT-3-Powered Game to Generate Sexually Explicit Material Involving Children
Latitude's GPT-3-powered game AI Dungeon was reportedly abused by some players who manipulated its AI to generate sexually explicit stories involving children.
How French welfare services are creating ‘robo-debt’
A French welfare office using software to automatically evaluate cases incorrectly notified a woman receiving benefits that she owed €542.
Tesla on Autopilot Crashed Parked Michigan Police Car on Interstate
A Tesla Model Y on Autopilot collided with a parked Michigan State Police (MSP) car which had its emergency lights on, in Eaton County, Michigan, although no one was injured.
Amazon's Monitoring System Allegedly Pushed Delivery Drivers to Prioritize Speed over Safety, Leading to Crash
A lawsuit cited Amazon as liable in a crash involving its delivery driver, alleging that Amazon’s AI-powered driver monitoring system pushed drivers to prioritize speed over safety.
Dartmouth's Geisel School of Medicine Reportedly Used Canvas Activity Logs to Accuse Students of Cheating
Dartmouth's Geisel School of Medicine reportedly used Canvas learning-management activity logs and an internal analysis process to investigate remote-exam cheating during the 2020–2021 academic year. Seventeen medical students were charged after the school inferred that they accessed course materials during exams, but students and outside technical reviewers said automated Canvas activity may have been misread as intentional misconduct. Dartmouth later dropped the charges.
Amazon Allegedly Forced Deployment of AI-Powered Cameras on Delivery Drivers
Amazon delivery drivers were allegedly forced to consent to algorithmic collection and processing of their location, movement, and biometric data through AI-powered cameras, or be dismissed.
Year-long AI Surveillance Pilot in Two South Australian Aged Care Facilities Reportedly Overwhelmed Staff with False Positives
Between March 2021 and March 2022, an AI-enabled video and audio monitoring system was trialed in two South Australian aged care facilities. According to an independent audit commissioned by South Australia Health, the system produced over 12,000 false alerts, overwhelming staff and contributing to at least one missed real incident. The report concluded the system did not reach accuracy levels acceptable to staff or management during the pilot.
Facebook's Automated Tools Failed to Adequately Remove Hate Speech, Violence, and Incitement
Facebook's automated moderation tools were shown by internal documents performing incomparably to human moderators, and accounting for only a small fraction of hate speech, violence, and incitement content removal.
TikTok’s "For You" Allegedly Pushed Fatal “Blackout” Challenge Videos to Two Young Girls
TikTok’s recommendation algorithm was alleged in a lawsuit to have intentionally and repeatedly pushed videos of the “blackout” challenge onto children’s feeds, incentivizing their participation which ultimately resulted in the death of two young girls.
Manufacturing Robot Failure Caused Factory Worker's Death in India
A sensor snag resulted in an automotive parts factory robot falling on a factory worker in India
Facebook’s and Twitter's Automated Content Moderation Reportedly Failed to Effectively Enforce Violation Rules for Small Language Groups
Facebook's and Twitter were not able to sufficiently moderate content of small language groups such as the Balkan languages using AI, allegedly due to the lack of investment in human moderation and difficulty in AI-solution design for the languages.
California's Algorithm Considered ZIP Codes in Vaccine Distribution, Allegedly Excluding Low-Income Neighborhoods and Communities of Color
California's vaccine-distribution algorithm used ZIP codes as opposed to census tracts in its decision-making, which critics said undermined equity and access for vulnerable communities who are largely low-income, underserved neighborhoods with low Healthy Places Index scores.
Facebook's Advertisement Moderation System Allegedly Routinely Misidentified Adaptive Fashion Products as Medical Equipment and Blocked Their Sellers
Facebook platforms' automated ad moderation system falsely classified adaptive fashion products as medical and health care products and services, resulting in regular bans and appeals faced by their retailers.