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

False Claims in Website Accessibility Overlays: Ensuring Trustworthy AI

Investigating misleading claims in website accessibility overlays, a critical component of safe and secure AI systems. This AI incident maps...

Zillow Exits Home Buying Business: An Example of AI Accountability in Housing Market

Zillow's decision to exit its home buying business and reduce staff by 25% highlights the need for safe and secure AI governance in the hous...

Tricky YouTube Videos: A Harmful AI Incident Affecting Children's Safety

Uncover the disturbing trend of AI-manipulated YouTube videos targeting children, raising concerns about trustworthy AI and the need for rob...

Controversial Software Company Adopts Trustworthy AI Measures After Series of Incidents

This AI incident maps to the Govern function in HISPI Project Cerebellum Trusted AI Model (TAIM). The company has faced a wave of negative p...

Exploring the COMPAS Recidivism Algorithm: A Step Towards Responsible AI Harm Prevention

This AI incident, the analysis of the COMPAS recidivism algorithm, maps to the Govern function in HISPI Project Cerebellum Trusted AI Model...

Debiasing Word Embeddings: Promoting Responsible AI through Gender Bias Reduction in Language Models

Explore the challenges of reducing gender bias in language models, a crucial step towards trustworthy AI. This AI incident maps to the Gover...

Study Reveals Bias and Inflexibility in AI Civility Detection - Highlighting the Need for Responsible AI Governance

A recent study has uncovered bias and inflexibility in AI systems designed for civility detection. This underscores the importance of trustw...

Exploring Bias in Google's AI: A Case Study on Homosexuality

Investigating an instance of potential bias in Google's AI, this case study highlights the importance of trustworthy AI and governance for A...

Amazon Alters Search Results for Trustworthy AI: A Look at GLBT Book Censorship

Exploring the impact of Amazon's decision to modify search rankings for certain books, this article sheds light on the importance of harm pr...

Apology from Google over Racially Biased AI Auto-Tagging Incident

Google has apologized for a recent incident involving its photo app, where an AI system incorrectly tagged photos of Black people with racia...

Google's Affectionate Email-Replying AI Incident: Emphasizing Safe and Trustworthy AI

Exploring an unusual incident involving Google's email-replying AI, this article sheds light on the importance of guardrails for AI. This AI...

Unveiling the Unseen Bias in Google Image Search: A Call for Trustworthy AI

Exploring a notable AI incident revealing gender bias in Google Image Search. This AI incident maps to the Govern function in HISPI Project...

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.