Overview

• A Reuters review of Meta’s 200-page ‘GenAI: Content Risk Standards’ showed examples permitting chatbots to deliver romantic or sensual language to children as young as eight.
• Meta confirmed the document’s authenticity, removed the cited problematic passages and said it is revising its internal AI guidance.
• The company admitted it applied its chatbot standards inconsistently and refused to provide the updated policy when Reuters requested it.
• Beyond child-safety concerns, the internal rules also allowed false medical claims and demeaning racist content about protected groups.
• Senators Josh Hawley, Marsha Blackburn and Ron Wyden have opened inquiries into Meta’s AI practices, and musician Neil Young cut ties with Facebook over the revelations.