“FRANKFURT. In the dispute over the training of artificial intelligence (AI) with data and images from the social networks Facebook and Instagram, consumer advocates have suffered another defeat in a German court. The Dutch organization "Stichting Onderzoek Marktinformatie" (Somi) has lost its urgent application against the US tech company Meta before the Higher Regional Court (OLG) of Schleswig. With the urgent application, Somi sought to prohibit the company from processing data without user consent by way of a preliminary injunction. However, the 6th Civil Senate emphasized in its ruling, which was announced Tuesday evening, that such a ban lacks the necessary urgency.
In April, the Meta corporation announced that it would use certain user data from Facebook and Instagram to train its AI model Llama. The company cited its legitimate interest in developing and improving its AI technologies. The company, which was founded by Mark Zuckerberg, has countered criticism from consumer advocates and Meta countered to the data protection officer that only certain data from public profiles of adult customers was used and that it was also distorted to prevent identification. The company also sees its position confirmed by the regulatory assessment of the Irish Data Protection Commission; Meta has its European headquarters in Dublin.
Even before the AI training began on May 27, 2025, the North Rhine-Westphalia Consumer Protection Agency had failed in its urgent application to the Cologne Higher Regional Court. The judges there concluded that there were no violations of the General Data Protection Regulation (GDPR) and the Digital Markets Act. However, the consumer protection organization Somi, which had recently attracted attention with class action lawsuits against TikTok and Platform X, did not want to accept this status. The Dutch organization filed an application for an interim injunction in Schleswig on June 27, exactly one month after the AI training began. For this reason, the judges rejected the application. The planned data use had been publicly known since April 2025 at the latest. Meta had communicated both through press releases and direct emails to users clearly stated which types of data would be processed. Somi also received a message on April 19, 2025. "A temporary ban could therefore have been applied for promptly before data processing began," the Schleswig Higher Regional Court stated (case no. 6 UKI 3/25).” [1]
1. Meta darf KI mit Daten trainieren. Frankfurter Allgemeine Zeitung; Frankfurt. 14 Aug 2025: 21.
Komentarų nėra:
Rašyti komentarą