DevelopingSafety
Meta adds LLM detection for ads that covertly point to child sexual abuse material
Meta says new LLM detection flags ads that signpost child exploitation content, alongside destination analysis, expanded sweeps, a red-teaming AI agent and repeat-offender detection. It reports 33.2 million pieces actioned globally in January to June 2026, 97% found proactively.
- HEAT
- OUTLETS
- 11
- NEW IN 6H
- +11
- FIRST SEEN
- LAST SIGNAL
Primary source
Metaabout.fb.com/news/2026/10/measures-weve-put-in-place-to-fight-child-exploitation/Coverage · 14 articles, oldest first
Wednesday, October 7, 2026