Meta, the parent company of Facebook and Instagram, has made a surprising announcement regarding its plans to train its artificial intelligence (AI) models using public content and user conversations from the European Union. This decision, revealed on Monday, April 14, 2025, signifies a significant shift in the company's strategy concerning data regulation in Europe.
Historically, Meta has been cautious in navigating the stringent data protection laws in Europe. However, it appears that the company, founded by Mark Zuckerberg, is now exploring new avenues to pursue its ambition of achieving dominance in the AI sector, directly competing with OpenAI's ChatGPT and Google's Bard.
Under this new policy, Meta AI will be trained on a variety of public content, including Instagram photo captions and Facebook comments. Nevertheless, Meta has emphasized that users retain the right to opt-out if they do not wish for their data to be utilized in AI training.
"This initiative will enhance our AI's ability to understand and reflect the culture, language, and history of Europe, ultimately providing more relevant services," the company stated in its official announcement.
Meta also clarified that WhatsApp will not be affected by this change, nor will private messages or accounts belonging to users under the age of 18.
For those who wish to opt-out, the company claims that the opt-out form is easily accessible and user-friendly, and Meta has committed to respecting all objections submitted, both past and future.
This development follows the delayed launch of Meta AI in Europe, which was postponed for over a year due to strict regulations surrounding new technologies, including data protection and digital market laws. When it was finally released in March 2025, Meta confirmed that the initial version had not been trained using data from European users.
With this latest decision, Meta joins Google and OpenAI, which have already adopted a similar approach in the European region.