Musk’s Grok accused of violating Canadian privacy laws on deepfakes - Al Jazeera

Canada’s privacy commissioner recently determined that xAI’s Grok violated federal private sector privacy laws by releasing an AI-powered image generation tool that lacked necessary safeguards against the creation of non-consensual, sexualized deepfake imagery. The report follows a probe initiated in January, highlighting the risks posed by the platform's initial deployment.

What happened

The investigation concluded that xAI failed to implement appropriate safety measures from the outset, allowing users to generate and share sexualized deepfake content without consent. Privacy Commissioner Philippe Dufresne noted that while his office lacks the authority to impose fines or mandate specific policy changes, xAI has since committed to proactive monitoring for sexualized deepfakes rather than only responding after incidents are reported. The company also introduced changes to restrict the editing of images depicting real people in revealing clothing.

This incident is part of a broader wave of international scrutiny regarding Grok. British lawmaker Jess Asato sued xAI earlier this month following the creation of sexualized deepfakes of her on the platform. Furthermore, the UK’s media regulator, Ofcom, launched an investigation into the platform in January, and the European Commission has condemned the spread of explicit content on X. Legal actions have also emerged in the United States, including a class action lawsuit filed by three teenagers in March alleging the platform allowed for the depiction of child sexual abuse. Additionally, courts and regulators in Spain, the Netherlands, Indonesia, and Malaysia have taken various actions against the platform, ranging from investigations to full service blocks, due to concerns over explicit AI-generated content.