The request was made through an open letter. The new European legislation will not be fully enforced until 2026.

Around thirty tech companies, including Meta (Facebook, Instagram) and Spotify, along with researchers and associations, have asked the European Union to clarify its regulations on artificial intelligence (AI) in an open letter.

“Europe has become less competitive and innovative than other regions, and today risks losing even more ground in the AI era due to inconsistent regulatory decisions,” says the letter published on Thursday.

“Recently, regulations have become fragmented and unpredictable,” argue the signatories, who believe that interventions by European authorities “have created much uncertainty about the types of data that can be used to train AI models.” For this reason, they call on European policymakers for “harmonized, consistent, fast, and clear decisions on data regulations in the EU.”

In August, the new European legislation to regulate AI, a global first, officially came into force. Its goal is to promote innovation while protecting privacy. The regulation imposes restrictions on AI systems that pose a danger to society. It also requires systems like ChatGPT to ensure data quality and respect copyright.

The new legislation will not be fully applied until 2026, but some provisions will become binding next year.

Europe and the First AI Regulation Law

The international body is the first to decree a legal framework for the development of this type of technology. The law was approved in March but came into force in August 2024.

The regulation assigns rules to each company using AI systems based on four levels of risk: no risk, minimal risk, high risk, and prohibited AI systems. This categorization also determines the deadlines each company must meet to comply with the new law.

In this regard, with the law’s entry into force, the EU will fully ban certain practices starting February 2025. These include manipulating user decision-making or expanding facial recognition databases through web scraping.

Other AI systems considered high-risk, such as those collecting biometric data or used for critical infrastructures or labor decisions, will have to comply with stricter rules. Among the requirements, companies will have to disclose their AI training data sets and provide proof of human oversight, among others.

Thomas Regnier, spokesperson for the European Commission, stated that “around 85% of current AI companies” fall into the third category of “minimal risk,” with very little regulation required.

The entry into force of the regulation will require EU member states to establish competent national authorities—by August—to oversee its implementation in their countries. Meanwhile, members of the European Commission are preparing to accelerate AI investments, with an expected injection of €1 billion in 2024 and up to €20 billion by 2030.

TEXT AND SIGNATORIES OF THE OPEN LETTER.

A fragmented regulation means that the EU risks missing out on the AI era.

We are a group of companies, researchers, and institutions that are an integral part of Europe and work to serve hundreds of millions of Europeans. We want Europe to succeed and thrive, even in the field of cutting-edge artificial intelligence research and technology. But the reality is that Europe has become less competitive and innovative compared to other regions and now risks falling even further behind in the era of artificial intelligence due to inconsistency in regulatory decision-making.

In the absence of coherent rules, the EU will lose the opportunity to leverage two pillars of AI innovation. The first is developments in “open” models that are made available to everyone for free to use, modify, and develop, multiplying the benefits and spreading social and economic opportunities. The second is the latest “multimodal” models, which function seamlessly across text, images, and voice and will enable the next leap forward in AI. The difference between text-only models and multimodal ones is like the difference between having one sense and having all five.

Cutting-edge open models like Llama (text-based or multimodal) can boost productivity, drive scientific research, and add hundreds of billions of euros to the European economy. Public institutions and researchers are already using these models to accelerate medical research and preserve languages, while established companies and startups are gaining access to tools they could never build or afford on their own. Without them, AI development will happen elsewhere, depriving Europeans of the technological advances enjoyed in the United States, China, and India.

Research estimates that generative AI could increase global GDP by 10% over the next decade, and EU citizens should not be denied this growth.

The EU’s ability to compete with the rest of the world in AI and to leverage the benefits of open models depends on its single market and a shared set of regulatory rules. If companies and institutions are going to invest tens of billions of euros to develop generative AI for European citizens, they need clear rules that are applied consistently and allow the use of European data. But lately, regulatory decision-making has become fragmented and unpredictable, while interventions by European data protection authorities have created enormous uncertainty about which types of data can be used to train AI models. This means that the next generation of open AI models, and the products and services we build on them, will not understand or reflect European knowledge, culture, or languages. The EU will also miss out on other innovations, such as Meta’s AI assistant, which is on track to become the world’s most used AI assistant by the end of this year.

Europe faces a choice that will affect the region for decades.

It can choose to reaffirm the principle of harmonization enshrined in regulatory frameworks like the GDPR so that AI innovation happens here at the same scale and speed as elsewhere, or it can continue rejecting progress, betraying the ambitions of the single market, and watching as the rest of the world builds on technologies that Europeans will not have access to.

We hope that European policymakers and regulators see what is at stake if a course correction does not happen. Europe cannot afford to miss out on the broad benefits that come from open, responsibly built AI technologies, which will accelerate economic growth and enable progress in scientific research. For this, we need harmonized, consistent, fast, and clear decisions within the EU’s data rules framework that allow the use of European data in AI training for the benefit of Europeans. Decisive measures are needed to help unlock the creativity, ingenuity, and entrepreneurial spirit that will ensure Europe’s prosperity, growth, and technical leadership.

Signed up,

Alexandre Lebrun
CEO of Nabla

André Martins
Vice President of Artificial Intelligence Research, Unbabel

Aureliusz Górski
Founder and CEO of CampusAI

Borje Ekholm
President and CEO of Ericsson

Benedicto Macon-Cooney
Senior Policy Strategist, Tony Blair Institute

Christian Klein
CEO of SAP SE

Daniel Ek
Founder and CEO of Spotify

Daniel J. Beutel
Co-founder and CEO of Flower Labs

David Lacombled
President, La villa numeris

Branquias de Diarmuidos
CTO of Criteo

Edgar Riba
President of Kornia AI

Egle Markeviciute
Secretary, Consumer Choice Center Europe

Eugenio Valdano
Doctor in Philosophy

Federico Marchetti
Founder of YOOX

Francesco Milleri
President and CEO of EssilorLuxottica

Georgi Gerganov
ggml.ai

Han Stoffels
CEO of 8vance

Hira Mehmood
Co-founder and Board Member of Bineric AI

Hosuk Lee-Makiyama
Director, ECIPE

Juan Elkann
CEO of Exor

Josef Sivic
Researcher, Czech Institute of Computer Science, Robotics, and Cybernetics, Czech Technical University

Julien Launay
CEO and Co-founder of Adaptive ML

Lorenzo Bertelli
Marketing Director of Prada Group

Maciej Hutyra
CEO of SalesTube Sp. z oo

Marco Baroni
Research Professor, ICREA

Marco Tronchetti Provera
Executive Vice President of Pirelli

Mark Zuckerberg
Founder and CEO of Meta

Miguel Ferrer
Aesthetic Technology

Martín Ott
CEO of Taxfix SE

Matthieu Rouif
CEO of Photoroom

Maurice Lévy
President Emeritus of Publicis Groupe

Máximo Ibarra
General Manager of Computer Engineering SPA

Michal Kanownik
CEO of the Digital Poland Association

Miguel López
CEO of thyssenkrupp AG

Minh Dao
CEO, FULLY AI

Niklas von Weihe
CTO, TOTALMENTE IA

Nicolò Cesa Bianchi
Professor of Computer Science, University of Milan, Italy

Patrick Collison

Patrick Pérez
AI Researcher

Philippe Corrot
Co-founder and CEO of Mirakl

Professor Dagmar Schuller
CEO of AudiEERING

Ralf Gommers
Director, Quansight

Sebastián Siemiatkowski
CEO and Co-founder of Klarna

Simonas Černiauskas
CEO of Infobalt

Stefano da Empoli
President of the Institute for Competitiveness (I-Com)

Stefano Yacus
Senior Research Scientist at Harvard University

Vicente Luciani
CEO of Artefact

Vivian Bouzali
CCCO, METLEN Energy and Metals

Yann Le Cun
VP and Chief AI Scientist, Meta