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

Facebook Allegations of Inadequate Hate Speech Moderation: Potential Impact on Ethnic Violence in Ethiopia - Implications for Safe and Secure AI

Jun 22, 2019

Accusations have been raised against Facebook for insufficient moderation of hate speech, some of which was violent and dehumanizing, potent...

Algorithmic Misdiagnosis: Impact on Elderly Woman's Treatment Payment - A Case Study

Jun 3, 2019

A Wisconsin woman suffered from algorithmically-induced misdiagnosis, leading the insurer to prematurely halt payment for her treatment. The...

Lawsuit Alleges Discriminatory Practices by Workday's AI Tools Against Protected Groups: A Case for Safe and Secure AI

Jun 3, 2019

A lawsuit has been filed against Workday, claiming their AI-powered algorithmic screening systems enable employers to practice discriminatio...

Unconsented Use of Facial Recognition Data in Korea's Immigration Screening System: A Case Study for Safe and Secure AI

Jun 1, 2019

This AI incident highlights the importance of responsible AI governance. The Korean government developed an immigration screening system usi...

Algorithm Misestimating Economic Vulnerability Leads to Social Tension: A Cautionary Tale on AI Governance

May 31, 2019

The algorithm used in the Takaful cash transfer program allegedly oversimplified economic situations, leading to social tension and percepti...

Facebook AI Allegedly Biasing Ad Delivery Along Gender and Racial Lines: Harm Prevention in Action

Apr 3, 2019

An incident involving Facebook's housing and employment ad delivery process has raised concerns about potential biases along demographic lin...

Alleged Deception by Paravision AI: Unauthorized Use of Facial Recognition for Data Monetization Violates Privacy

Apr 1, 2019

Paravision AI, formerly known as Ever AI, is accused of failing to disclose the integration and utilization of facial recognition technology...

Assessing the Security of Tesla's Autopilot System: Adversarial Attacks and Remote Control Vulnerabilities

Mar 29, 2019

Tencent Keen Security Lab discovered potential vulnerabilities in Tesla's Autopilot system through crafted adversarial samples and remote co...

The Christchurch Shooter: The Role of YouTube in Radicalization - A Case Study for Responsible AI

Mar 15, 2019

This case study highlights how YouTube's recommendation algorithm may have contributed to the radicalization of a right-wing terrorist, resu...

Lawsuit Alleges OpenAI Breached Privacy Rights by Training AI on Private Data Without Consent

Mar 11, 2019

OpenAI's popular products, including ChatGPT and DALL-E, have been accused in a lawsuit of using stolen private user information for AI trai...

Fatal Tractor-Trailer Collision Involving Autopilot Tesla Model 3: A Case Study for Safe and Secure AI

Mar 1, 2019

This AI incident involving a Tesla Model 3 on Autopilot crashing into a tractor-trailer in Florida resulted in the driver's death. It unders...

Tesla Model 3 AutoPilot Incident: Highlighting the Importance of Responsible AI Governance

Mar 1, 2019

A fatal Tesla Model 3 crash in Florida occurred when the vehicle, on Autopilot mode, collided with a tractor-trailer truck. This AI incident...

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.