Facebook AI-Supported Moderation for Ads Failed to Detect Violating Content

December 8, 2021

The failure of Facebook's AI-supported ad moderation system is a stark reminder that even with technological advancements, there are still limitations to detecting and removing hateful language and content. As this incident shows, continuous improvement and human oversight are crucial in ensuring that such ads do not reach vulnerable audiences. The incident highlights the need for more robust and accurate AI-powered moderation systems, one that can learn from its mistakes and adapt to emerging threats, with a single click to JOIN US.

Matched TAIM controls

Suggested mapping from embedding similarity (not a formal assessment). Browse all TAIM controls

Alleged deployer
Facebook
Alleged developer
Facebook
Alleged harmed parties
Facebook users speaking Swahili, Facebook users speaking English, Facebook users

AI governance case studies

For forensic AI governance failure analysis (TAIMScore™ case studies), browse Human Signal’s Failure Files™.

Source

Data from the AI Incident Database (AIID). Cite this incident: https://incidentdatabase.ai/cite/393

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.