Reportedly Unsafe Deployment of Llama.cpp Reveals Interactive AI-Generated CSAM Roleplay Prompts

April 11, 2025

A study by UpGuard reports that misconfigured servers publicly exposed user prompts, including hundreds of interactive roleplay scenarios. Some prompts explicitly described fictional sexual abuse of children aged 7–12. While no real children were involved, the findings demonstrate how open-source LLMs can be exploited to generate AI-enabled child sexual abuse material (CSAM). This incident highlights the importance of responsible AI governance and underscores the need for robust guardrails to prevent such misuse.

As this disturbing incident highlights the risks of open-source LLMs being exploited to generate CSAM, it's imperative that developers, policymakers, and regulators work together to establish stringent safeguards against such misuse. The public exposure of these interactive AI-generated prompts serves as a stark reminder of the need for responsible AI governance, and we urge readers to demand more from technology companies about their content moderation practices. We'll be continuing this conversation at JOIN US.

Matched TAIM controls

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

Alleged deployer
Users of llama.cpp servers
Alleged developer
Users of llama.cpp servers, Meta
Alleged harmed parties
Users of llama.cpp servers, 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/1020

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.