A PYMNTS Company

X Halts Use of EU Data for AI Training Amid Legal Scrutiny

By  |  August 8, 2024

Social media giant X, formerly known as Twitter, has agreed to temporarily cease using personal data from European Union (EU) users for training its artificial intelligence (AI) systems, as confirmed in an Irish court on Thursday. The move comes amid growing concerns from regulators about the platform’s data practices.

Ireland’s Data Protection Commission (DPC), which oversees the operations of many major U.S. tech companies in Europe due to their EU headquarters being based in Ireland, requested earlier this week that X be ordered to stop or limit the processing of user data for AI development. According to Reuters, this legal action reflects increasing regulatory pressure on tech firms over their handling of personal data, especially in the realm of AI.

X, owned by billionaire Elon Musk, has stated that it offers users the option to control whether their public posts can be utilized by its AI chatbot, Grok. However, this opt-out feature requires users to manually adjust their privacy settings by unticking a specific box. The company asserted that this option is available to all users, enabling them to protect their data if they choose.

Read more: Elon Musk’s X Files Antitrust Suit Against Global Advertising Alliance

Despite these assurances, Judge Leonie Reynolds highlighted in court that X had already begun using EU users’ data for AI training as early as May 7, but did not provide the opt-out option until July 16. Furthermore, she pointed out that the feature was not immediately accessible to all users, raising concerns about the adequacy of X’s data protection measures.

The scrutiny of X’s data practices follows a similar situation involving Meta Platforms, another U.S. tech giant. In June, Meta decided to postpone the European launch of its AI models after the Irish DPC advised against proceeding with the plan. This trend signals a broader unease among European regulators about the use of personal data for AI purposes without sufficient user consent.

Source: Reuters