Meta has received permission from the European Data Protection Authority to train artificial intelligence (AI) models with public data from users on its platforms, including Facebook, Instagram, WhatsApp and Messenger.
The decision was announced on April 14 in a blog post of the company. Messages, reactions and other public interactions of adult users may now be used to improve Meta’s AI systems.
Meta states that it is essential to train AI models with a wide range of data, so that they better match cultural and linguistic diversity within Europe. The company mentions examples such as dialects, spoken language and even the diverse ways in which humor and sarcasm are expressed in different countries.
Privacy is guaranteed: Users can make choices
Although Meta now has permission to use public data, the privacy of users is guaranteed. The company has indicated that private messages between users, as well as data from EU users under the age of 18, will not be used for AI training. Moreover, users can easily indicate that they do not want their data to be used for the training of AI models. Meta offers a form that is available both via the app and e-mail.
The company should have postponed this step earlier after the Privacy Organization None of Your Business had filed objections against the use of personal data for AI purposes. The Irish Committee for Data Protection (IDPC) had asked Meta to temporarily pause the rollout for an evaluation.
Approval of European supervisor for AI training
Meta has since received approval from the European Commission for Data Protection. According to the regulator, Meta’s approach meets the requirements of European privacy legislation.
The company emphasizes that similar methods are already being applied in other regions. Other major technology companies, such as Google and OpenAI, also use public data from European users to train their AI systems.
Research into other tech giants and new EU legislation
Despite this approval, there are still concerns about compliance with privacy legislation by other tech companies. In September 2024, the Irish regulator started a cross-border investigation into Google, to check whether the company complies with the EU rules in the development of AI. X received similar criticism, and the company decided to adjust its policy by no longer using EU user data for the training of his AI-Chatbot Grok.
In August 2024, the EU introduced the AI Act, which provides clear guidelines for the development and use of artificial intelligence, with emphasis on data quality, security and privacy.
Source: https://newsbit.nl/meta-krijgt-goedkeuring-voor-ai-training-met-publieke-gegevens-van-europese-gebruikers/
