Values Statement: We believe AI should cause no harm, but enhance the quality of human life, by proactively adopting our AI Governance framework.

Evidence-based Transparent For governance

AI Incidents

TAIM catalog Map Data source & citation

Google Apologizes for Racist Auto-Tag Incident in Photo App

In an incident that raises concerns about AI governance, Google's photo app automatically tagged images of black people with offensive terms...

AI Analysis Determines Criminality Based on Facial Traits

A new AI program uses facial analysis to predict criminal behavior, raising concerns about responsible AI governance and the potential for h...

Making AI Fairer: Test Your Skills in Our Courtroom Algorithm Game

Explore the intricacies of AI fairness in a unique, interactive courtroom setting. This game highlights the challenges and solutions in crea...

The Incident of Tay (Chatbot): A Case Study in Responsible AI

In 2016, the chatbot Tay, developed by Microsoft's research team, made headlines for all the wrong reasons. Designed to learn from and engag...

Fatality in an Unfortunate Incident Involving a Self-Driving Uber Vehicle

In this critical incident, a pedestrian lost their life after being struck by a self-driving Uber vehicle. The event serves as a stark remin...

The Perils of Franken-Algorithms: Navigating the Risk of Unpredictable AI

Exploring the dangers of unforeseen consequences in complex AI systems, akin to Frankenstein's monster. Understanding and implementing respo...

The Unforeseen Consequences of Autonomous Vehicles: Uber's Self-Driving Car and the Jaywalker Dilemma

An exploration into the tragic incident involving Uber's self-driving car, highlighting the AI's inability to comprehend jaywalkers. The eve...

Autonomous Vehicle Incident: Fatal Collision of Self-Driving Uber Car in Arizona

In March 2018, an autonomous Uber vehicle struck and killed a pedestrian in Tempe, Arizona, highlighting the urgent need for robust safety m...

Algorithmic Bias: Unveiling the Impact on Kidney Transplants for Black Patients

Recent research has uncovered a concerning incident involving an algorithm that has inadvertently affected the allocation of kidney transpla...

Examining the Controversies Behind Efficient Spam Filters

Spam filters, designed to protect inboxes from unsolicited emails, have been a staple of email services for decades. However, their efficien...

Government's Algorithm Performance in Question: An Examination of Responsible AI

A recent incident involving the use of algorithms by the government has sparked debate over their implementation, safety, and transparency....

Algorithm Misidentifying Jewish Infants in Strollers

A recent AI-powered image recognition algorithm has been found to misidentify Jewish infants in strollers as adults. This incident underscor...

Data source

Incident data is from the AI Incident Database (AIID).

When citing the database as a whole, please use:

McGregor, S. (2021) Preventing Repeated Real World AI Failures by Cataloging Incidents: The AI Incident Database. In Proceedings of the Thirty-Third Annual Conference on Innovative Applications of Artificial Intelligence (IAAI-21). Virtual Conference.

Pre-print on arXiv · Database snapshots & citation guide

We use weekly snapshots of the AIID for stable reference. For the official suggested citation of a specific incident, use the “Cite this incident” link on each incident page.