Unlock the Editor’s Digest for free
Roula Khalaf, Editor of the FT, selects her favourite stories in this weekly newsletter.
Europe’s data protection watchdog is “seeking clarity” on a decision by Elon Musk’s X to allow users’ data to automatically be fed into his artificial intelligence start-up xAI, placing fresh regulatory scrutiny on the social media platform.
X users discovered on Friday that they had been ‘opted-in’ to having their posts to the site, as well as their interactions with its Grok chatbot, be used for “training and fine-tuning” xAI’s systems.
The move was made without first obtaining users’ explicit consent for data sharing. The setting can only be changed on the desktop version of X, so users are currently unable to opt out via its mobile apps.
Ireland’s Data Protection Commission, the regulator responsible for holding internet companies to EU privacy law, said that it had been engaging with X “for months” over its plans to employ its users’ data to create AI systems. The regulator had sent questions to the company as recently as Thursday about “transparency for users” among other issues, it said.
“It took us by surprise that they were rolling this out,” a spokesperson for the watchdog said, adding that the agency had followed up on Friday with further questions “seeking clarity” from X on the matter.
Privacy experts have questioned whether X’s move might breach the EU’s General Data Protection Regulation rules, which require companies that collect or use personal data to first obtain the individual’s consent and disclose their reason for doing so. If the Irish regulator opens a GDPR probe, X could face fines or penalties.
Meta last month paused a plan to train its AI on data from its Facebook and Instagram platforms in Europe following a request from the Irish DPC around GDPR compliance concerns. This marked “a step backwards for European innovation, competition in AI development”, Meta said.
Musk bought X for $44bn in 2022, and has faced growing scrutiny from regulators globally as he has cut back its staff and moderation capabilities, and brought in sudden product changes. The EU Commission is already investigating the company for alleged transparency and moderation issues in breach of its Digital Services Act.
“All X users have the ability to control whether their public posts can be used to train Grok, the AI search assistant,” X’s safety account wrote on Friday. “This option is in addition to your existing controls over whether your interactions, inputs, and results related to Grok can be utilised. This setting is available on the web platform and will soon be rolled out on mobile.” X did not respond to a request for comment.
The rollout comes as Musk is aiming to gain an edge on rivals including OpenAI, Anthropic and Google, whose most advanced models are currently “orders of magnitude” ahead of xAI’s, the billionaire has said. Access to X’s massive repository of user data would provide an important advantage to xAI. Grok, meanwhile, is accessible to X premium subscribers.
Musk’s network of companies has become increasingly interlinked. Engineers and other senior staff have moved between his businesses, and the serial entrepreneur is seeking approval from Tesla’s board to invest up to $5bn into xAI.
Some investors in xAI have cited its synergies with other Musk entities as a critical advantage, though others have raised concerns about potential conflicts.