Anthropic Safety Filters Block Review of Aztec History Book

Patrick McKenzie posted a screenshot showing an LLM attempting to transcribe or review Clendinnen’s Aztecs. The model encountered an automated content policy block attributed to Anthropic safety filters due to graphic violence descriptions in the historical text. The reply presents this as an instance of safety layers preventing access to the material. The evidence packet contains only the post, the image, and the generated headline summarizing the event. No independent confirmation or additional context on the model version or exact prompt appears in the source lines.