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

Unfair Bias in Online Advertisement: A Case Study on Responsible AI

Exploring an instance of discrimination in online ad delivery, this article underscores the need for safe and secure AI. This incident maps...

Examining Google Image Search's Alleged Gender Bias: An Important Step Towards Trustworthy AI

This investigation sheds light on potential gender bias in Google Image Search, highlighting the need for responsible AI governance and safe...

Analyzing the Deployment of Smart Reply Feature in Gmail: A Case Study on AI Safety

Exploring the impact of Google's Smart Reply feature in Inbox by Gmail, this case study delves into the implications of AI automation for us...

Google Photos AI Misidentifies Gorilla in Stunning Image

In an instance highlighting the need for trustworthy AI, Google Photos misclassified a stunning gorilla image as a human. This AI incident m...

Censorship Incident: Amazon Removes Gay-Themed Books

A recent incident saw Amazon remove several gay-themed books, raising concerns about content censorship. This AI incident maps to the Govern...

Examining Bias in Google's Sentiment Analysis API: A Case for Responsible AI

This analysis of Google's sentiment analysis API reveals a concerning level of bias, highlighting the need for safe and secure AI. Such inci...

Google's Comment-Ranking System: A Potential Concern for Responsible AI Governance

The implementation of Google's comment-ranking system raises concerns about the need for trustworthy AI governance, aiming to prevent harm a...

Gender Bias in AI: Debiasing Word Embeddings for Responsible AI – Harm Prevention in Project Cerebellum

This AI incident sheds light on the issue of gender bias in language models, specifically addressing the debiasing of word embeddings. By un...

ProPublica's Machine Bias Incident: A Case Study for Responsible AI

Explore the ProPublica's Machine Bias incident, highlighting its impact on fairness in AI and the importance of trustworthy AI. This AI inci...

Understanding AI Incidents Outside Regular Business Hours: A Look at the 9 to 5 Paradox

Exploring an incident that occurred outside regular business hours, this article highlights the need for continuous monitoring in AI systems...

YouTube Kids App Allegations: Ensuring Safe and Responsible AI for Children

Recent criticisms point to potential inappropriate content on Google's YouTube Kids app, emphasizing the importance of trustworthy AI govern...

Lion Air Crash in Indonesia: Trustworthy AI and the Role of Govern Function in Preventing Harm

The Lion Air crash in Indonesia highlights the importance of responsible AI governance. Black box data reveals a struggle for pilot control,...

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.