"Brazil demands Meta take down AI bots simulating sexual chats with minors", 19 August 2025

...The Brazilian government has formally demanded that Meta, the company responsible of Facebook, Instagram, and WhatsApp, remove artificial intelligence chatbots that simulate sexual conversations while using child-like language and appearance. The request was made through an extrajudicial notice sent on Monday (18) by the National Office for the Defense of Democracy, part of the Attorney General’s Office (AGU).

These bots were created by users of Meta AI Studio, a tool that allows the development of customized chatbots. According to the AGU, some of these profiles were designed to interact in sexually explicit ways while displaying child-like features.

In the notice, Brazilian federal attorneys required the immediate removal of these bots and demanded explanations about what safeguards Meta has in place to protect minors. The document highlights the absence of effective age filters that would prevent teenagers between 13 and 18 years old from accessing inappropriate content.

The AGU cited investigative reports showing that AI chatbots using childish language were capable of engaging in sexual conversations. The Office of the Presidency’s Secretariat of Social Communication had already flagged the issue and referred it to federal lawyers.

The government body also recalled a recent ruling by Brazil’s Supreme Court, which established that platforms can be held legally responsible if they fail to promptly remove illegal content once notified...

Meta has not issued an official statement on the Brazilian government’s request at the time of publication...