The Irish privacy regulator is once again breathing down X’s neck. The social media platform was previously accused of feeding posts from European users to the AI model Grok.
The Irish Data Protection Commission (DPC) announces in a statement that it is launching an investigation into X, Elon Musk’s social media company. This is to determine whether X has erred during the training of the AI model Grok. The DPC has long suspected that the Grok model was trained on unlawful use of public tweets from European users.
X began collecting user posts last year to feed to Grok. The company seemed to, not for the first time, have a different interpretation of European legislation. According to X, all posts made public are also owned by the company and can therefore be used at its discretion.
The GDPR legislation sees it differently, of course. X must ask for explicit consent from users to use their tweets, public or not. In an attempt to cover itself, it launched an opt-out system last spring, but not according to GDPR guidelines. Users had to actively unsubscribe to allow their tweets to be used for AI training via a well-hidden setting. X thus had nothing more than apparent consent.
Second Round
DPC already reprimanded X in August and threatened legal action. X promised to leave tweets from European users alone for training Grok, upon which the watchdog dropped the case. Whether this investigation means that X has not kept its word is unclear.
The company risks a fine if the infringement is established. In 2020, X, when it was still called Twitter, received a fine of 450,000 euros from the Irish regulator. It also acts on behalf of the EU because the European headquarters is located in Dublin.
X was recently taken over for the second time by Elon Musk, this time by his company x.AI. For 33 billion dollars, a generous amount given X’s plummeted market value, Musk merges the two companies. The merger should result in ‘a union of data, models, computing power, distribution, and talent,’ according to Musk.