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)
-
Security Robot Drowns Itself in a Fountain AIID ↗
A Knightscope K5 security robot ran itself into a water fountain in Washington, DC.
-
Deepfake Obama Introduction of Deepfakes AIID ↗
University of Washington researchers made a deepfake of Obama, followed by Jordan Peele
-
Houston ISD's EVAAS Teacher-Evaluation System Reportedly Put Teachers' Jobs at Risk Through Unverifiable Scores AIID ↗
From 2011 to 2015, Houston Independent School District used SAS Institute’s Education Value-Added Assessment System (EVAAS) to rate teacher effectiveness from student test-score growth. Teachers alleged that proprietary, unverifiable EVAAS scores contributed to terminations and contract nonrenewals. On May 4, 2017, a federal judge allowed due-process claims over the system to proceed.
-
Nissan's "Automatic Emergency Braking" False Positives Posed Traffic Risks to Drivers AIID ↗
Nissan's Automatic Emergency Braking (AEB) feature was reported in a series of complaints for false positives and abrupt braking behaviors, endangering car occupants and traffic participants.
-
High-Toxicity Assessed on Text Involving Women and Minority Groups AIID ↗
Google's Perspective API, which assigns a toxicity score to online text, seems to award higher toxicity scores to content involving non-white, male, Christian, heterosexual phrases.
-
AI incident: sexual content (Amazon, Dec 2016) AIID ↗
An AI Incident Database record involving sexual content, dated 30 December 2016. Risk classification, related incidents and the laws on record for this risk area.
-
Reinforcement Learning Reward Functions in Video Games AIID ↗
OpenAI published a post about its findings when using Universe, a software for measuring and training AI agents to conduct reinforcement learning experiments, showing that the AI agent did not act in the way intended to complete a videogame.
-
Poor Performance of Tesla Factory Robots AIID ↗
The goal of manufacturing 2,500 Tesla Model 3's per week was falling short by 500 cars/week, and employees had to be "borrowed" from Panasonic in a shared factory to help hand-assemble lithium batteries for Tesla.
-
LinkedIn Search Prefers Male Names AIID ↗
An investigation by The Seattle Times in 2016 found a gender bias in LinkedIn's search engine.
-
Common Biases of Vector Embeddings AIID ↗
Researchers from Boston University and Microsoft Research, New England demonstrated gender bias in the most common techniques used to embed words for natural language processing (NLP).
-
Security Robot Rolls Over Child in Mall AIID ↗
On July 7, 2016, a Knightscope K5 autonomous security robot collided with a 16-month old boy while patrolling the Stanford Shopping Center in Palo Alto, CA.
-
A Collection of Tesla Autopilot-Involved Crashes AIID ↗
Multiple unrelated car accidents result in varying levels of harm have been occurred while a Tesla's autopilot was in use.
-
COMPAS Algorithm Reportedly Performs Poorly in Crime Recidivism Prediction AIID ↗
Correctional Offender Management Profiling for Alternative Sanctions (COMPAS), a recidivism risk-assessment algorithmic tool used in the judicial system to assess likelihood of defendants' recidivism, is found to be less accurate than random untrained human evaluators.
-
Northpointe Risk Models AIID ↗
An algorithm developed by Northpointe and used in the penal system is two times more likely to incorrectly label a black person as a high-risk re-offender and is two times more likely to incorrectly label a white person as low-risk for reoffense according to a ProPublica review.
-
Is Pokémon Go racist? How the app may be redlining communities of color AIID ↗
Through a crowdsourcing social media campaign in 2016, several journalists and researchers demonstrated that augmented reality locations in the popular smartphone game Pokemon Go were more likely to be in white neighborhoods.
-
Google Waymo Vehicle and Bus Allegedly Collide in California AIID ↗
On February 14, 2016, a Google autonomous test vehicle partially responsible for a low-speed collision with a bus on El Camino Real in Google’s hometown of Mountain View, CA.
-
Arkansas's Opaque Algorithm to Allocate Health Care Excessively Cut Down Hours for Beneficiaries AIID ↗
Beneficiaries of the Arkansas Department of Human Services (DHS)'s Medicaid waiver program were allocated excessively fewer hours of caretaker visit via an algorithm deployed to boost efficiency, which reportedly contained errors and whose outputs varied wildly despite small input changes.
-
Amazon Alexa Responding to Environmental Inputs AIID ↗
There are multiple reports of Amazon Alexa products (Echo, Echo Dot) reacting and acting upon unintended stimulus, usually from television commercials or news reporter's voices.
-
Predictive Policing Biases of PredPol AIID ↗
Predictive policing algorithms meant to aid law enforcement by predicting future crime show signs of biased output.
-
Collection of Robotic Surgery Malfunctions AIID ↗
Study on database reports of robotic surgery malfunctions (8,061), including those ending in injury (1,391) and death (144), between 2000 and 2013.
-
Google and Delphi Self-Driving Prototypes Allegedly Involved in Near-Miss on San Antonio Road, Palo Alto AIID ↗
A Delphi self-driving Audi Q5 prototype reportedly aborted a lane change after a Google self-driving Lexus RX400h allegedly cut in front of it on San Antonio Road in Palo Alto, California. According to Delphi executive John Absmeier, the Delphi car "took appropriate action" to avoid collision. No crash or injuries occurred.
-
Facebook's 'Tag Suggestions' Allegedly Stored Biometric Data without User Consent AIID ↗
Facebook's initial version of the its Tag Suggestions feature where users were offered suggestions about the identity of people's faces in photos allegedly stored biometric data without consent, violating the Illinois Biometric Information Privacy Act.
-
Images of Black People Labeled as Gorillas AIID ↗
Google Photos image processing software mistakenly labelled a black couple as "gorillas."
-
Google’s YouTube Kids App Presents Inappropriate Content AIID ↗
YouTube’s content filtering and recommendation algorithms exposed children to disturbing and inappropriate videos.
-
Google's Face Grouping Allegedly Collected and Analyzed Users' Facial Structure without Consent, Violated BIPA AIID ↗
A class-action lawsuit alleged Google failing to provide notice, obtain informed written consent, or publish data retention policies about the collection, storage, and analysis of its face-grouping feature in Google Photos, which violated Illinois Biometric Information Privacy Act (BIPA).
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.