Meta, the parent company of Facebook and Instagram, has announced that it will pause its plans to train its artificial intelligence (AI) models with content shared by adults on these platforms across the European Union and the European Economic Area. This decision was made in response to concerns raised by citizens, civil society, and legal groups such as Noyb about the lack of explicit consent from users, complicated opt-out processes, and the potential use of images without consent. The Irish data protection authority (DPC), where Meta has its European headquarters, welcomed the decision and expressed gratitude to Meta for taking into account the concerns of citizens.
The decision to train AI models using user-generated content faced significant opposition from a variety of stakeholders. Activists like Max Schrems emphasized the need for a firm resolution on this matter. On the other hand, Meta expressed disappointment with DPC’s decision, arguing that it hinders innovation and competitiveness in AI development in Europe. The company stated that it complies with data protection regulations and only uses publicly shared content to build language models. The Privacy Director at Meta emphasized that the social network does not use private messages or content from minors for AI training purposes, and that users voluntarily make public their data for this purpose.
Despite this decision to suspend AI training with user content, discussions on data privacy and innovation in AI continue. The interaction between social media platforms, data protection authorities, and user rights will likely shape the future of AI development and data protection policies in Europe.
