'EU changes course on Ai and data'. The appeal of companies and research
Meta, in an open letter published, brings together dozens of leading European companies to ask EU legislators to promote new, less restrictive rules on Ai
5' min read
5' min read
Meta, in an open letter published in major newspapers, brings together several dozen of Europe's leading companies to ask lawmakers to promote new, less restrictive rules on artificial intelligence. 'A change of course,' they write, 'in order not to be excluded from the great benefits of an open technology capable of accelerating economic growth and research'.
Among the signatories of the document addressed to European legislators and regulators - including EssilorLuxottica, Prada, Pirelli, Exor Group and Spotify - are also names from research and companies such as Engineering, Ericsson, Nicolò Cesa-Bianchi (University of Milan), Eugenio Valdano, PhD (Sorbonne/Inserm.).
Already at the end of August, Mark Zuckerberg, who financed the publication of the letter, had intervened in person in the weekly magazine The Economist against the EU's restrictive policies, describing them as risky at a time in history when losing the AI train could jeopardise the future.
This message has been picked up and relaunched by an important part of the European business community, which writes: 'Europe is faced with a decision that will have consequences for the continent for decades. It can choose to reaffirm the principle of harmonisation enshrined in legal frameworks such as the GDPR and offer a modern interpretation of its provisions that still respects its core values, thus allowing innovation in AI to develop here with the same scope and speed as in other regions of the world. Or, it can continue to reject progress, contradict the ambitions of the single market, and stand by while the rest of the world develops technologies that European citizens will not have access to'.
The role of Meta and open source systems
The letter, in particular, focuses on the goodness and attractiveness for Europe of open source models - Meta's Llama is among them - and multimodal models, which represent an evolution of text chatbots. We are talking about software that marks an evolutionary leap, as it allows images, video and audio to be processed. As if to say, in addition to 'talking', these large language models are also acquiring other 'senses'. A few months ago, Meta had decided not to offer its next multimodal artificial intelligence model and future versions to customers in the European Union, citing a lack of clarity on the part of regulators in Brussels. The Irish Data Protection Commission (DPC), Ireland's data protection authority, had in July asked Meta to delay training large language models (LLMs) using public content shared by European adults on Facebook and Instagram.This decision came as a surprise to Meta, as initially the DPC, which regulates GDPR enforcement in the EU as many global tech giants, including Meta, Google and Apple, have their European headquarters in Ireland, had given the green light to the company's AI in Europe.
