Meta announced new plans to train its generative AI models using public content from users based in the European Union. Beginning this week, adults using Meta’s platforms will be informed through in-app notifications and emails about the data usage. The information explains that public posts and comments will help teach the AI to better understand European languages, cultures and history. Users are provided with an easy-to-access form if they wish to object to this data usage. The move highlights transparency regarding how user interactions contribute to AI model improvement.
The company clarified that private communications such as messages with friends and family are not included in the training process. Additionally, public content created by EU users under the age of 18 will not be used. This distinction is designed to address privacy concerns while supporting the broader goal of developing AI systems that effectively reflect the cultural and linguistic diversity of Europe.
Meta’s approach follows its recent launch of Meta AI across Europe which marked the company’s entry into the region with its chat functionality. By integrating AI into widely used platforms such as Facebook, Instagram, WhatsApp and Messenger, the initiative seeks to create tools that are both accessible and relevant to European users. This practice is consistent with methods used by other technology companies in the region including Google and OpenAI, who have similarly utilized public data to train their AI models.
In complying with European laws and regulations, Meta’s decision comes after previous delays while legal guidelines were clarified. The company references the opinion provided by the European Data Protection Board, confirming that their original approach met legal obligations. Meta continues to engage with the proper regulatory bodies to ensure the responsible deployment of generative AI technology in Europe.
