Report from Politico
In Brief – Ireland’s data watchdog has opened an investigation into allegations that X’s Grok generative AI tool created potentially harmful, non-consensual intimate images of European individuals, including of children, improperly using their personal data. The Irish Data Protection Commission said that the investigation is being undertaken through its authority as the lead enforcement authority for X’s compliance with the EU’s General Data Protection Regulation (GDPR). Deputy Commissioner Graham Doyle said the inquiry will assess X’s compliance with GDPR obligations in the operations of its AI image generation tools and follows weeks of engagement with the company. The GDPR-based action adds to the queue of investigations by government officials in markets across the globe scrutinizing X’s integration of Grok into the social media platform with functionality that allowed users to create and post realistic AI-generated images that could include depictions of women, and sometime apparently children, in revealing clothing and sexualized poses.
Context – Last summer, X integrated the Grok Imagine image tool with a “spicy mode” that permitted the creation of sexually suggestive content. Although the tool quickly generated complaints from civil society advocates, there was little public notice until Reuters reported in January on what it called a “flood of nearly nude images of real people” created by Grok and circulating on X in markets around the world. In Europe, where AI image generators are not yet directly regulated by the AI Act, the GDPR probe joins one by the EU Commission under the Digital Services Act, a criminal investigation by French prosecutors, and potentially a criminal investigation in Spain as well. Regulators in the UK, Brazil and California are also engaged. Although Elon Musk initially dismissed the charges, the company eventually responded to the pressure with what Musk described as a local law policy, with the AI chatbot capable of generating only images that he said are legal in each region. The changes did not satisfy critics who claimed that “sexualized” images could still be created by Grok on its standalone app.