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 Content Moderators Call for Improved Working Conditions Due to Inadequate AI-Based Content Moderation: A Harm Prevention Issue

Apr 1, 2020

Content moderators and employees at Facebook are advocating for better working conditions, raising concerns over the inadequacy of the curre...

Kargu-2 Drone Allegedly Used for Remote Surveillance and Attacks: A Case of Potential AI Misuse

Mar 27, 2020

This incident involves a Turkish-made Kargu-2 drone, which is equipped with computer vision, being potentially misused in Libya by forces li...

Voice Recognition Systems Struggle with Accuracy for Black Speakers: A Discussion on Responsible AI

Mar 23, 2020

A recent study revealed that voice recognition tools from major tech companies, including Apple, Amazon, Google, IBM, and Microsoft, exhibit...

Facebook's Alleged Anti-Spam Bug: Impacts on COVID-19 Information Sharing - Harm Prevention through Responsible AI

Mar 17, 2020

Recent reports claim that Facebook blocked legitimate posts about the coronavirus pandemic, potentially due to a bug in its anti-spam system...

Unavailability of Video: An Incident Illustrating the Importance of Safe and Secure AI

Mar 16, 2020

This AI incident, involving the unavailability of a video, highlights the critical role of trustworthy AI in our daily lives. It maps to the...

Harm Prevention: Social Media Content Removal Hindering Investigative Efforts - Safe and Secure AI

Mar 16, 2020

Incident involving alleged hindrance of investigative efforts due to automated removal of violating social media content, such as terrorism,...

Canada: Universities' AI Proctoring Tools Allegedly Failed Consent Requirements - Harm Prevention in Remote Learning

Mar 9, 2020

The use of AI proctoring tools during remote exams by Canadian universities reportedly violated individual consent, as biometric data collec...

Alleged Racial Bias Reinforcement in TikTok's Recommendation Algorithm: A Case Study of AI Incident Mapping to the Govern Function in HISPI Project Cerebellum Trusted AI Model (TAIM)

Feb 24, 2020

TikTok's 'Suggested Accounts' algorithm allegedly displayed racial bias, despite not using race or profile photos for recommendations. This...

Self-Driving Shuttle Incident Highlights Importance of Safe and Secure AI: EasyMile Unexpectedly Stops Mid-Route, Injuring a Passenger

Feb 20, 2020

The unexpected stopping of a self-driving shuttle by EasyMile in Columbus' Linden neighborhood raises awareness about the need for trustwort...

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.