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

TikTok Anorexia Videos Evade 'Pro-Ana' Ban through Spelling Tactics

This AI incident showcases a challenging aspect of content moderation in social media platforms, emphasizing the need for robust and respons...

Account Deletion of Latina Trans Woman on TikTok Highlights Need for Responsible AI

This AI incident involving account deletion on TikTok underscores the importance of trustworthy AI and its impact on marginalized communitie...

AI-Controlled Mall Robot Malfunctions on Escalator, Causes Passenger Injuries: A Case for Safe and Secure AI

This AI incident involving a mall robot malfunctioning and causing harm to passengers underscores the need for responsible AI governance. Th...

Analyzing a College Admissions AI: The Intersection of Fairness, Bias, and Responsible AI

Exploring an incident involving a college admissions algorithm: understanding its impact on fairness, bias mitigation, and the role of respo...

The Impact of Brand Safety Technology on News Funding: A Case Study in Unintended Consequences

Explore the unforeseen consequences of brand safety technology, which can lead to news defunding. This incident maps to the Govern function...

Israeli Farmer vs AI Algorithm: A Case Study in Responsible AI

An incident between an Israeli farmer and an algorithm showcases the importance of trustworthy AI and safe, secure AI governance. This AI in...

Privacy Concerns Lead to University of Illinois Suspending Use of Remote-Testing Software

The University of Illinois has temporarily halted the use of its remote-testing software following student complaints about potential privac...

Analyzing Algorithmic Bias in E-commerce Vaccine Misinformation: A Step Towards Responsible AI

This AI incident highlights the importance of scrutinizing algorithmically curated content for misleading vaccine information on e-commerce...

Exam Monitoring Software Bias Against BIPOC Students at University of Toronto Highlights Importance of Responsible AI

An incident involving bias in exam monitoring software affecting BIPOC students at the University of Toronto underscores the need for trustw...

Beverly Hills Cop Using Sublime's 'Santeria': AI Analysis on Potential Privacy Breach

This intriguing incident raises questions about privacy in live-streaming and its impact on responsible AI. This AI incident maps to the Gov...

Understanding AI Bias in Social Media Ad Approval: A Case Study on Fashion Ads

In this article, we explore an incident where fashion ads were repeatedly rejected by Facebook's AI system. This case study highlights the c...

Struggles in Maintaining Trustworthy AI: Addressing Content Violations on Facebook and Twitter

This AI incident highlights the challenges in ensuring safe and secure AI, specifically within content moderation on popular platforms like...

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.