Meta will use public posts to train the IA also in Europe. Private messages and minors’ data will remain excluded.
Meta, the company that controls the social Facebook, Instagram, Threads and the Whatsapp messaging service, has announced the start of training its generative artificial intelligence models using the public content shared by adult users on its platforms in the European Union. The initiative, which follows the recent launch of Meta Ai in Europe, was designed to improve the ability of the digital assistant to understand and reflect the different cultures, languages and traditions present on the continent.
Methods of use of data and respect for privacy
Meta clarified from the beginning that private messages exchanged with friends and family will not be used, nor public data from account account under 18 years. In particular, the contents intended for training include posts, images, videos and comments made public, in addition to interactions with destination AI, the chatbot integrated in the main messaging applications such as Messenger and Instagram Direct. WhatsApp, however, remains excluded from this specific operation, as the consent for the use of data for training has already been requested separately.
A global approach with focus on the European market
The approach adopted by Meta is not an exclusive novelty for Europe. In fact, the giant has already implemented similar methodologies to train its generative artificial intelligence models in other regions of the world. Following the example of other leaders in the sector, such as Google and Openai, Meta underlines the need to have diversified and qualitatively high datasets, capable of representing the peculiar linguistic and cultural characteristics of the various countries.
The company also highlighted how the suspension of the training of its linguistic models, adopted last year pending regulatory clarifications, represented a moment of important reflection. After welcoming the opinion of the European Data Protection Board (EDPB) and started a profitable dialogue with the Irish Data Protection Commission (IDPC), a destination is said to be ready to bring the benefits of generative artificial intelligence to European users, in full compliance with current legislation.
With this move, Meta aims to strengthen its positioning in the field of generative artificial intelligence, a rapidly evolving sector where the quality and diversity of training data are fundamental elements. The goal is to create increasingly performing models that can adapt to the cultural and linguistic specificities of a varied continent such as the European one. This represents a significant step in the company’s innovation strategy, which intends to remain competitive globally, while maintaining a strong commitment to the protection of privacy and respect for European regulations.
Notifications and the possibility of dissent
To ensure transparency and compliance with European legislation, Meta will send personalized notifications to users residing in the EU, both through apps and email. These communications will illustrate in detail which types of data will be used and how this information will contribute to perfecting artificial intelligence systems and improving the overall experience of users. A fundamental aspect of the initiative concerns the possibility of dissent: notifications will include a direct link to the form (which you also find at this link) which will allow users to oppose the use of their data for training the IA models.
