Numerous complaints X's AI chatbot learns with user data without being asked
Twitter successor X lets its AI chatbot Grok learn from users' posts without being asked. The company is now facing complaints under the European General Data Protection Regulation.
No time? blue News summarizes for you
- The data protection organization Noyb has filed complaints in eight EU countries against X (formerly Twitter) for the unsolicited use of user data.
- Noyb criticizes the Irish data protection authority DPC because its complaint against X only covers secondary aspects of the case.
- X users were not informed about the use of their data and the corresponding data protection settings can only be changed in the web version.
The European data protection organization Noyb has filed complaints against Elon Musk's online platform X in eight EU countries. The "urgent proceedings" under the European General Data Protection Regulation (GDPR) on behalf of affected EU citizens concern the X chatbot Grok, whose artificial intelligence is trained by default with contributions from users.
As X users were not informed in advance about the use of their data for AI training and were not asked for permission, the Irish data protection authority DPC, which is responsible for the Twitter successor X in Europe, filed a complaint against the network last Tuesday. However, this does not go far enough for the data protection association Noyb because the complaint only deals with secondary aspects of the case.
Max Schrems, Chairman of Noyb, said: "Companies that interact directly with users simply have to ask them a yes/no question before using their data. They do this regularly for many other things, so it would definitely be possible for AI training too."
Urgency required
In view of the fact that X had already started processing people's data for its AI technology, Noyb had requested an "urgency procedure" under Article 66 GDPR in Austria, Belgium, France, Greece, Ireland, Italy, the Netherlands and Spain. This provision allows data protection authorities to issue provisional orders in such situations and to take an EU-wide decision via the European Data Protection Board.
The unsolicited use of X data and the changed data protection settings were noticed two weeks ago by the X user @EasyBakedOven. The checkbox for Grok's permission to use public X posts in addition to direct interactions with the chatbot was automatically checked for everyone.
The setting can only be changed in the web version of X; it is currently not displayed at all in the smartphone app. X announced that this would change soon.
Meta postponed plans - X implements them
In June, the Facebook group Meta postponed its plans to use public posts from users in Europe to train its AI models indefinitely following pressure from Irish data protection authorities. Previously, Meta had been criticized for not providing explicit user consent, but only the option to object to the use of the data. X is now proceeding as Meta intended.
Grok is intended to compete with other AI chatbots such as the pioneer ChatGPT from OpenAI or Claude from Anthropic. The software is not being developed directly by X, but by the company xAI, which is also owned by Musk. He bought Twitter in the fall of 2022 for around 44 billion dollars and renamed the service X.
dpa
