Meta will not offer its new generative AI models in Europe due to its “unpredictable regulatory environment”
The company decides to exclude the EU in its deployment of the latest version of Llama a month after Apple announced that some of its features would not be available in the Old Continent
JUL 18, 2024 - The Country of Spain
Meta goes one step further in its fight with Brussels. The company has announced today that it will not deploy its new generative artificial intelligence (AI) models in Europe, which are capable of generating content from written instructions. The decision is a response to what Menlo Park considers excessive interference by the EU in its industry through the complex European regulatory framework.
“In the coming months we will launch a multimodal Llama model, but not in the EU due to the unpredictable nature of the European regulatory environment,” a spokesperson for the technology giant said in a statement. The new version of Llama will be capable of processing videos, audios, texts and images and can be used from the mobile phone or from the Meta Ray-Ban smart glasses .
Although the Llama model will be released with an open license, European companies will not be able to use it. In the same way, companies from outside the EU that incorporate this tool will not be able to offer services related to it in European territory. What Meta will launch in the EU, according to company sources, is a text-only version of its Llama 3 model
The decision of the company founded by Mark Zuckerberg comes a few days after Meta itself renounced using the data of its Facebook or Instagram users to train its AI models . This movement, in turn, was prompted by the investigation opened in this regard by the data protection authority of Ireland, the country in which Meta has its European headquarters.
Apple, for its part, said a month ago that it will not deploy in Europe all the new AI-based functionalities it is working on, considering that some of its features could clash with European regulation.
One more disagreement
The multinational held meetings with members of the Commission to explain what its new data collection policy would be like to train its AI. It also told users how they could opt out of having their data used in those workouts, a process that many commentators have called extremely cumbersome. All of this would have influenced the decision not to deploy its advanced AI models in European territory.
Just two weeks ago, the European Commission accused Meta of violating the Digital Markets Act (DMA) with its “payment or consent” model, by which users of Facebook and Instagram, the star social networks of the company, who do not want to be bombarded with personalized advertising can avoid it by paying a monthly fee. Meta can now present allegations and the Commission will resolve in 12 months the result of its investigation, which could lead to a fine of 10% of its annual turnover (amounting to approximately 13,000 million euros).
The European Data Protection Board ruled at the end of last year that Meta could no longer process personal data of its users to serve them advertising tailored to their profile.
In recent years, the EU has built a pioneering regulatory architecture in the world to protect its citizens from the growing power of big technology companies. The first step was the entry into force in 2018 of the General Data Protection Regulation (GDPR), a rule that requires users to be informed about what data is being taken from them and for what purposes.
The next leg, made up of two regulations (the Digital Services Law, DSA, and the Digital Markets Law, DMA ), establishes transparency and accountability mechanisms specifically aimed at large technology companies, Meta among them. The third component of this defense shield is the Artificial Intelligence Regulation , approved at the beginning of the year and which will come into force on August 1.
Today is not the first order that Meta has launched in his fight with European legislators. In February 2022, the technology company hinted that it would leave Europe (that is, European citizens would not be able to use Facebook, Instagram or WhatsApp) if the legislation was not favorable to it. At that time, one of the major disputes between Big Tech and the EU was about to be decided: whether or not the former can host the data of European citizens on US territory, which would imply that they are not subject to the restrictive regulations. European. Finally, the US and the EU closed an agreement the following year . And Meta didn't leave.