X attempts to evade Brazil's investigation into AI deepfakes, and the Brazilian government pushes back
X Tried to Sidestep Brazil's Inquiry on AI Deepfakes. The Government Just Pushed Back.
Brazilian regulators rejected X's claim that Grok is a separate entity and that @Grok is just an ordinary user account, and escalated demands that X implement technical safeguards or face daily fines. According to the Center for Countering Digital Hate, Grok generated more than 3 million sexualized images in 11 days starting December 29, of which 23,000 were suspected to involve minors. In a letter to Brazil's Ministry of Justice's Secretariat for Digital Rights, X acknowledged that its policies had failed, but said X Brazil, X Corp, and xAI LLC, which operates Grok, are legally independent of one another. On January 20, Brazil's Federal Public Ministry, National Data Protection Authority, and the Ministry of Justice's Consumer Secretariat jointly required Grok to prevent the generation of sexualized images without consent and to establish processes for identifying and removing them. A Reuters investigation on February 3 found that after X claimed safeguards had been added, Grok still generated 29 sexualized images in 43 attempts.
Brazilian regulators rejected X's attempt to shift responsibility by treating Grok as an ordinary user, and changes to its platform liability provisions also affect generative AI regulation.