This concentration of power makes European governments uncomfortable: European companies become the downstream customers of the future, importing the latest services and technologies in exchange for money and data sent across the Atlantic to the west. These concerns have taken on new urgency, partly because some in Brussels sense a widening gulf in values and beliefs between Silicon Valley and the EU’s average citizen and elected representatives, and partly because AI looms large in the collective imagination as the driver of the next tech revolution.
Concerns about Europe falling behind in AI predate ChatGPT. In 2018, the European Commission released an AI plan calling for “Made in Europe” AI that could compete with the US and China. But beyond a desire to have some control over the shape of the technology, the operational definition of AI sovereignty has been pretty hazy. “For some, it means we need to band together to take on big tech,” says Daniel Mügge, a professor of political arithmetic at the University of Amsterdam who studies EU tech policy. “For others, it means that big tech is fine as long as it’s European, so let’s hurry up and make it happen.”
These competing priorities are beginning to complicate EU regulation. The EU’s AI law, passed by the European Parliament in March and expected to become law this summer, focuses on regulating potential harms and privacy concerns surrounding the technology. But some member states, notably France, have made clear during negotiations over the law that they worry the regulations will hamstring their own AI startups, which it hopes will become a European alternative to OpenAI.
Speaking ahead of a UK summit on AI safety last November, French Finance Minister Bruno Le Maire said Europe needed to “innovate before regulating” and that the continent needed “an AI-savvy European actor.” The final text of the AI bill includes a commitment to make the EU “a leader in the adoption of trustworthy AI.”
“Italy, Germany and France thought at the last minute, ‘We need to be a bit more lenient with European companies on the foundational model,'” Mügge says. “It boils down to this idea that Europe needs European AI. I think people have since realised that this is a bit harder than they would have liked.”
Sahlin, who recently traveled to European capitals, including meeting with policymakers in Brussels, said Europe has several ingredients it needs to compete: To thrive in AI, he said, it needs data, computing power, talent and capital.
Data is relatively widely available and Europe has AI talent, but it can sometimes struggle to retain it, Sarin added.
To mobilize more computing power, the EU is investing in high-performance computing resources, creating a pan-European network of high-performance computing facilities, and providing access to supercomputers to start-ups through its “AI Factory” initiative.
Raising the capital needed to build large-scale AI projects and companies is also difficult, with a big gap between the United States and other countries. According to Stanford University’s AI Index report, private investment in U.S. AI companies will exceed $67 billion in 2023, more than 35 times the amount invested in Germany and France. According to Accel Partners research, the top seven private investment rounds in U.S. generative AI companies in 2023 will total $14 billion. The top seven in Europe totaled less than $1 billion.
