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

Incident Analysis: Facebook AI Moderator Mistakes Car Washes for Mass Shootings

This alarming AI incident highlights the need for robust safeguards in our AI systems. The Facebook moderator, tasked with filtering content...

Ethnicity Detection Algorithm Issue: Alibaba's Cloud Unit Under Scrutiny - Mapping to the Govern Function of HISPI Project Cerebellum Trusted AI Model

This incident involving Alibaba's ethnicity detection algorithm highlights the importance of responsible AI and governance. Stay informed ab...

Third of California Bar Exam Applicants Flagged for Suspected AI-Assisted Cheating: An Alarming Trend in the Race for Trustworthy AI

This alarming incident highlights the need for robust AI governance in high-stakes exams. The California Bar Exam flagged a third of applica...

TikTok Anorexia Content Evading Filters: A Concern for Responsible AI Governance

This AI incident highlights the need for safe and secure AI governance, particularly on social media platforms. TikTok anorexia videos have...

Account Deletion of a Latina Trans Woman on TikTok: An Examination of AI Bias in Content Moderation

This incident highlights the potential bias in AI-powered content moderation systems, raising concerns about responsible AI governance and h...

Unsafe AI Incident: Shopping Mall Robot Falls from Escalator, Endangering Passengers

A disturbing incident occurred in a shopping mall where an autonomous robot fell off an escalator, causing injuries to several passengers. T...

Exploring the Impact of an Admissions AI Algorithm: Lessons in Trustworthy AI

Dive into a critical analysis of an admissions algorithm, highlighting its implications on fairness and bias in AI systems. This incident ma...

The Impact of Ineffective Brand Safety on News Funding - A Cautionary Tale for Responsible AI

Explore the detrimental effects of inadequate brand safety technology on news funding, emphasizing the importance of trustworthy and safe AI...

Israeli Farmer Challenges Algorithm Decision-Making: A Case Study in AI Governance

Explore the clash between an Israeli farmer and an autonomous irrigation algorithm, highlighting the need for safe and secure AI. This incid...

AI Bias in Educational Software: A Case of Students of Color Being Misidentified

An AI incident involving educational software has been reported, where students of color are being flagged to teachers due to the software's...

Auditing E-commerce Platforms: Preventing AI-driven Spread of Vaccine Misinformation - A Step Towards Trustworthy AI

Investigating algorithmic curation practices on e-commerce platforms to combat vaccine misinformation, a critical aspect of safe and secure...

Exam Monitoring Software Disadvantages Affecting BIPOC Students at the University of Toronto: An Instance Highlighting the Need for Trustworthy AI

This incident underscores the disadvantages faced by BIPOC students in their academic journey due to an untrustworthy AI application. Such b...

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.