EU Sets Out Plan to Close the AI Gap with US and China

The European Commission on Wednesday issued a report outlining the 27-member bloc’s strategy to establish Europe as a “leading AI continent.” The plan is intended to close the gap in AI innovation and capacity with the U.S. and China, and includes 5 core initiatives: a €20 billion ($22.1 billion) commitment to jumpstart investment in computing infrastructure; action to ensure access to high-quality data for AI innovators; steps to spur adoption of AI technology in key strategic sectors; developing a strong AI talent base; and leveraging the strengths of the EU single market.

The 25-page plan (PDF) alludes to “streamlining existing data legislation to reduce complexity and administrative burden and to ensure that data governance structures are efficient and effective,” but does not outline specific steps. It also emphasizes that the AI Act “must be enforced” and that any initiative under the plan must “[take] into account applicable copyright legislation.”

To that end, the plan calls for establishing an AI Act Service Desk that will be a “central information hub on the AI Act,” and “provide straightforward and free access to information and guidance on the applicable regulatory framework.” The Service Desk will be housed within the EU AI Office and will serve as an interactive platform where “businesses and other stakeholders, including public authorities, will be able to ask questions, get answers and have access to technical tools to help them apply the AI Act.”

Read more: OpenAI’s Viral Studio Ghibli Moment Highlights AI Copyright Concerns

While the plan is ambitious, the EU is starting from well behind the U.S. and China in the race for AI supremacy. A study released this week by researchers at Stanford found that 40 “notable [i.e. influential] AI models” have been developed by companies or institutions in the U.S., 15 emerged from China, but only three were developed in Europe, all in France.

To catch up, the plan calls for creating a network of “AI Factories” throughout Europe. The Factories will serve as hubs for connecting universities, startups, businesses, public sector organizations, and financial stakeholders with “supercomputing centers to foster innovation. To date, 13 consortia spanning 17 member states have been selected to host the Factories member states.

The report also aims to develop “AI Gigafactories” that would each host at least 100,000 “advanced AI processors” (i.e. GPUs). The plan targets at least five Gigafactory centers, to be funded through public-private partnerships via the European High-Performance Computing Joint Undertaking (EuroHPC) established in 2018 and updated in 2024.

As part of the Factories initiative, the plan calls for creating Data Labs to ensure AI developers have access to “large volumes of high-quality data in health, energy or other sectors (always in compliance with the rules that apply to each data space).” The Labs will be established through a new Data Union Strategy that will be unveiled later this year, the report said. The labs are intended to create “Common European Data Spaces” but could also offer other services, such as cleaning and enriching datasets, developing standardized formats, and providing synthetic data, as well as fostering interoperability across sectors and borders. The labs could also offer “data-pooling services” to act as “trusted intermediaries” to allow companies to share data without running afoul of EU antitrust laws.