Author: Supantha Mukherjee and Paul Sandle
STOCKHOLM/LONDON (Reuters) – Elon Musk’s Grok chatbot is testing Europe’s ability to combat deepfakes and digital stripping of images online, although regulators scored a rare victory by forcing Musk’s xAI to curb the creation of pornographic images.
xAI said late Wednesday it had restricted image editing by Grok AI users after they created thousands of pornographic images of women and minors, alerting regulators around the world.
Musk initially laughed off the trend, and his ouster highlights the difficulty of regulating artificial intelligence tools that can cheaply and easily create explicit content. It is the latest clash between Europe and Musk after a spat over election interference, content moderation and free speech.
Many regulators are still busy crafting laws and rules to govern artificial intelligence, but questions remain about what constitutes nudity, how consent is defined and who bears responsibility: users or platforms.
“It’s really a gray area when it comes to the creation of nude images,” Angla Pändel, a Stockholm-based data protection and privacy lawyer at Mannheimer Swartling, told Reuters.
British regulator Ofcom, one of the most outspoken regulators on the issue, welcomed Musk’s move but said it would continue to investigate xAI over the Grok images.
“Our formal investigation remains ongoing,” a spokesman said. “We are working around the clock to make progress and find answers to the issues and the steps we are taking to address them.”
Officials say strongest enforcement is still needed
Earlier this month, Grok created hyper-realistic images of women on X that were manipulated to look like they were wearing tiny bikinis, posing in degrading poses, and even covered in injuries. Some minors were digitally stripped down to their swimsuits.
As late as Wednesday, Reuters discovered that the chatbot was still secretly producing pornographic images on demand. On Thursday, the situation appeared to be contained, at least in some areas.
Musk’s xAI says it’s blocking users from generating images of scantily clad people in “jurisdictions where it’s illegal.” It did not identify those jurisdictions.
In Malaysia and Indonesia, governments imposed temporary bans on Grok, while EU and UK regulators called the images illegal. Britain, France and Italy launched investigations but faced calls for tougher action.
“Strengthened enforcement under the Digital Services Act (DSA) is needed to stop apps and platforms that sexualize or nudity women and children,” said Christian Democratic MEP Nina Carberry, who called the latest move a “positive step.”
A spokesman for the European Commission said that if Grok’s changes are not effective, the Commission will still use the EU DSA’s full enforcement toolbox to target the platform.
Legal gray area, heavy burden on victims
Alexander Brown, a UK-based data protection lawyer at Simmons & Simmons, said the UK’s online safety bill would make the sharing of intimate images without consent, including AI-generated deepfakes, a “priority offence”.
“This means that X must take proactive and appropriate steps to prevent such content from appearing on its platform and promptly remove it when discovered,” he said.
In the most serious breaches, UK regulators can fine companies up to 10% of revenue or ask a court to order internet service providers to block the site.
For individuals, taking a platform to court is “a very difficult and burdensome process,” said Anders Bergsten, a lawyer at Manheimers Watling, citing the emotional toll on victims.
Deepfakes have been around for years, long before artificial intelligence applications, although they have been largely confined to the dark corners of the web. X’s publishing capabilities give Grok unprecedented influence.
“The ability to publish frictionlessly allows deepfakes to spread at scale,” said Carrie Goldberg, a U.S. attorney who works with victims of online harassment.
UK and Swedish laws make it illegal to share nude photos without consent. The UK is extending the law to include the production of such images.
According to the DSA, suspension of service is considered a last resort. Experts say the EU Artificial Intelligence Act also does not have any provisions for adult nude images, only transparency obligations for deepfakes.
Prime Minister Keir Starmer welcomed X’s move on Thursday but warned: “Freedom of speech is not freedom from consent. Images of young women are not public property and their safety is not up for debate.”
“If we need to further strengthen existing laws we are prepared to do so.”
(Reporting by Supantha Mukherjee in Stockholm and Sam Tobin in London; Editing by Adam Jourdan, Kenneth Li and Elaine Hardcastle)
