Child Sexual Abuse Material Taints Image Generators

December 20, 2023

The LAION-5B dataset, a widely used dataset for image-description pairs, was found to contain child sexual abuse material (CSAM). This discovery has serious implications for downstream models, which may produce CSAM imagery. As a result, organizations are now retraining their models and scrubbing the original dataset of such imagery. Meanwhile, researchers are working on developing more robust AI governance frameworks to prevent such incidents in the future.

To ensure that this type of incident is reported and addressed, we invite you to visit JOIN US and learn more about our efforts towards responsible AI development and harm prevention.

The HISPI Project Cerebellum TAIM has established a database to track and report such incidents, providing critical guardrails for AI. We encourage your participation in shaping the future of trustworthy AI.

Matched TAIM controls

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

Alleged deployer
Various people, Various organizations
Alleged developer
LAION
Alleged harmed parties
Various people, Various organizations, Minors, LAION, General public

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/624

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.