The decision of the administration of the social network X to open user publications for Elon Musk’s startup xAI for training artificial intelligence under the presumption of user consent may violate European standards for protecting the privacy of citizens, the Financial Times reports.
Image source: NoName_13 / pixabay.com
The day before, X users discovered that they had “consented” to have their posts on the site and correspondence with the Grok chatbot used to “train and fine-tune” xAI’s AI. In reality, users did not give prior explicit consent to this. You can revoke it only in the desktop web version of X, and there is no corresponding option in mobile applications yet.
Ireland’s Data Protection Commission (DPC), the regulator responsible for ensuring big tech companies comply with European privacy regulations, said it had been in discussions with X for months about the company’s plans to use user data to train AI. Just on Thursday, the department sent a number of questions to the company, some of which related to ensuring “transparency for users.” “When they started [collecting user data for AI training], it was a surprise to us,” a commission spokesman said. On Friday, she sent X follow-up questions to “clarify” the situation.
There is an opinion that administration X, through its actions, violated the norms of the General Data Protection Regulation (GDPR) in force in the EU. The document requires companies to disclose the reasons for their actions and obtain consent from users before collecting user data. If the Irish regulator begins an investigation into a possible violation of the GDPR, the social network may face punishment. In June, Meta✴ abandoned plans to train its own AI on posts from European Facebook✴ and Instagram✴ users due to the threat of violating GDPR, however, calling this measure “a step back for European innovation, competition in AI development.”
Image source: BoliviaInteligente / unsplash.com
Since purchasing Twitter (now X) in 2022, Elon Musk has faced scrutiny from regulators around the world for dramatically cutting staff, including moderators, and making other changes to the company. The social network is currently under investigation on suspicion of violating the European Digital Services Act due to problems with moderation and transparency.
«All X users have the ability to control whether their public posts can be used to train Grok, the AI search assistant. This option is in addition to your existing controls over whether all of your Grok-related interactions, inputs, and results can be used. This customization is available on the web platform and will soon be rolled out to mobile devices,” the X security account posted.
Musk is looking to catch up and release AI that is on par with OpenAI, Anthropic and Google, and user data resource X will give xAI an edge, although access to Grok is only available to the social network’s premium subscribers. Musk’s companies are increasingly integrating with each other – for example, he is now seeking approval from the Tesla board of directors to invest $5 billion in xAI. Some of the startup’s investors cite its synergies with the billionaire’s other projects as an important advantage, while others have raised concerns about potential conflicts of interest.