Artificial intelligence feeds on data. But how far can it draw from our digital lives? The answer takes a new turn in Europe. Meta has just obtained the approval of European regulators to cause its AI models on public content shared by its users. A decision that raises as much technological hopes as ethical questions.
Meta crosses a regulatory course: AI nourished by social networks
The green light given to Meta, who recently landed in France, is not a white check. The company will be able to use the publications, comments and requests addressed to its AI assistant on Facebook, Instagram, Whatsapp or Messenger. But beware: private messages between relatives and data from under 18s remain taboos. A crucial distinction, presented as a “guarantee” by Meta.
Why this choice? The firm argues that the linguistic and cultural diversity of Europe requires models capable of grasping local dialects, Nice sarcastic humor or the Berlin hyperlocal references. “Without this variety, AI would become an awkward tool, unable to understand the shades that make our platforms vibrate”,, she explains. An attractive argument, but which is not enough to appease fears.
European users nevertheless have an exit door: an opposition form, promised easy to access.
It remains to be seen whether this mechanism will be as visible as an ephemeral Instagram story … or drowned in the meanders of the account parameters. Meta claims to “empower” its users. A fragile balance between transparency and practicality.
A regulatory context in changing: between opportunities and challenges
This authorization does not fall from the sky. She closed an showdown started in July 2023, when the NGO None of Your Business had obtained the project freezing, denouncing a possible operation of retroactive personal data.
After months of negotiations, Meta convinced the European Data Protection Commission that its approach respected the legal framework. A turnaround that illustrates the tensions between innovation and private life.
Meta is not a lonely pioneer. Google and Openai have already used European data for their AI, while X (ex-Twitter) had to give up causing Grok on EU user information. The difference? Meta relies on legitimacy by example: “We follow established actors”she underlines. A skillful strategy, but which could feed a vicious circle of standardization.
In this moving landscape, the EU tries to put safeguards. His law on artificial intelligence, which entered into force in August 2024, supervised data quality, security and confidentiality. However, surveys are multiplying, like that aimed at Google for possible shortcomings in the development of its models. Proof that the current framework remains an open site, where each authorization creates a precedent to monitor.
In this moving landscape, the EU tries to put safeguards. Its AI Act, which entered into force in August 2024, supervises data quality, security and confidentiality. However, surveys are multiplying, like that aimed at Google for possible shortcomings in the development of its models. Proof that the current framework remains an open site, where each authorization creates a precedent to monitor – especially since Google also changes the rules for crypto pubs in Europe with Mica, illustrating the growing scope of digital regulations on several fronts.
Maximize your Cointribne experience with our 'Read to Earn' program! For each article you read, earn points and access exclusive rewards. Sign up now and start accumulating advantages.
