OTTAWA, June 11 (Reuters) - xAI's Grok chatbot violated Canadian privacy laws by launching an image-generation tool that allowed users to create and share non-consensual, sexualized deep fakes, an official watchdog said in a report on Thursday.
After Privacy Commissioner Philippe Dufresne launched a formal probe in January, xAI announced tweaks to prevent Grok from allowing the editing of images of real people in revealing clothing. Britain and Canada are two in a growing list of countries cracking down on explicit content generated by Grok.
"xAI violated Canada's federal private sector privacy law by launching the Grok AI-powered image generation tool without implementing appropriate safeguards from the outset," Dufresne told a press conference.
Dufresne, who does not have the power to impose fines or order policy changes, said he was encouraged by the changes xAI had made to reduce the chances of sexualized deep fakes.
xAI has committed to regularly monitoring for sexualized deep fakes before an incident is reported, and not just in response to incidents, he said.
xAI, the artificial intelligence company owned by Elon Musk, was not immediately available for comment.
The Canadian government released a digital safety bill on Wednesday that would ban social media for children under 16 and set up a digital regulator.
(Reporting by David Ljunggren; Editing by Maria Cheng and Mark Porter)
