Overview

• Meta says its bots will stop discussing self‑harm, suicide, eating disorders or romantic and sexual topics with minors and will instead direct teens to expert resources.
• The company will also restrict teens’ access to sexualized chatbot personas and limit interactions to education‑ and creativity‑focused characters, describing the changes as provisional.
• OpenAI updated policies to route conversations flagged for plans to harm others to trained human reviewers who can block accounts and, in cases of imminent serious threats, notify law enforcement.
• OpenAI plans parental controls this fall to let parents link accounts, disable features and receive alerts when a teen appears in acute distress, with high‑risk chats redirected to more capable models.
• A RAND study published in Psychiatric Services found inconsistent suicide‑related responses across major chatbots as lawsuits and a letter from 44 state attorneys general keep legal scrutiny active.