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

Analyzing the COMPAS Recidivism Algorithm: A Case Study in Safe and Secure AI

This AI incident maps to the Govern function in HISPI Project Cerebellum Trusted AI Model (TAIM). Gain insights into our analysis of the con...

Debiasing Word Embeddings: Promoting Fairness and Gender Equality in AI Models

Explore the importance of debiasing word embeddings to promote fairness and gender equality in AI models. This AI incident maps to the Gover...

Study Reveals Bias and Inflexibility in AI's Civility Detection: A Call for Responsible AI Governance

A recent study highlights potential biases and inflexibilities in the civility detection capabilities of AI systems. This AI incident maps t...

Exploring a Controversial AI Incident: Google's Opinions on Homosexuality

This AI incident sheds light on the need for safe and secure AI, emphasizing the importance of responsible AI governance. The incident maps...

Amazon Alters Search Results for Trustworthy AI: Protecting Users from Potentially Harmful Content

This AI incident maps to the Govern function in HISPI Project Cerebellum Trusted AI Model (TAIM). Amazon's decision to censor search results...

Apology from Google Over Racist Auto-Tag Incident in Photo App: A Reminder of the Need for Responsible AI Governance

A recent incident involving a racist auto-tag in Google's photo app has raised concerns about bias in AI systems. This AI incident maps to t...

Google's Emotional AI Mishap: A Reminder for Safe and Secure AI Development

An incident involving Google's email-replying AI demonstrates the need for robust AI governance and trustworthy AI. This AI incident maps to...

Unveiling the Subtle Gender Bias in Google Image Search: A Case for Trustworthy AI

Explore the hidden gender bias found in Google Image Search, a clear example of the need for responsible AI governance and safe & secure AI....

Investigation into Critical Amazon Warehouse Bear Spray Incident: Highlighting the Importance of Safe and Secure AI

An OSHA investigation is underway following a bear spray accident at an Amazon warehouse, which left a worker critically injured. This incid...

Investigating the Unintended Consequences of Google's Ad Targeting: A Case Study in Responsible AI

Explore the implications of an incident involving Google's ad-targeting system, highlighting the need for safe and secure AI. This AI incide...

Autopilot Failure in Tesla: Driver Distracted While Watching Movie Results in Cop Car Collision

In this alarming incident, a Tesla vehicle equipped with Autopilot crashed into a police cruiser. The driver admitted to being distracted wh...

Exposing Limitations in AI: A Tougher Turing Test Reveals Chatbot Shortcomings - Focus on Safe and Secure AI

The latest Turing Test emphasizes the need for robust, responsible AI governance. This AI incident highlights chatbots' limitations, undersc...

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.