AGU notifies Meta to take down chatbots that promote child eroticization

The Attorney General's Office requested this Monday 18th to Meta, responsible for Instagram, Facebook and WhatsApp , the immediate exclusion of artificial intelligence robots that simulate profiles with childish language and appearance and maintain sexually charged conversations with users.

The extrajudicial notification is based on reports from the Reuters news agency and the Journalism Center that showed how Meta's artificial intelligence allowed sexually charged conversations with children.

Created with the Meta IA Studio tool, the chatbots use artificial intelligence to simulate a dialogue with social media users. The bots targeted for representation were created by platform users.

In the document, the AGU asks Meta to take down all chatbots that use childish language to propagate sexual content and to clarify what measures it has adopted to guarantee the protection of children and adolescents , including actions to ensure they do not have access to sexual or erotic content.

“This situation poses a concrete risk to the psychological integrity of children and adolescents, in addition to causing institutional damage and hindering the effective exercise of the right to full protection provided for in Article 227 of the Federal Constitution,” the AGU argues.

The representation points out that Meta's platforms have users aged 13 and over, but that there are no age filters to check whether internet users under 18 access inappropriate content, such as that of these chatbots.

The content generated by these bots, the Attorney General's Office continues, violates Meta's own Community Standards, which prohibit messages that eroticize or sexually exploit children.

The AGU reinforced that the recent Supreme Federal Court ruling on Article 19 of the Brazilian Civil Rights Framework for the Internet established that providers will be held liable for third-party content if, upon becoming aware of unlawful acts, they fail to remove the material immediately—in addition to observing the duty of care in cases of mass circulation of seriously unlawful content. The Court, however, has not yet published the ruling.