Mario Draghi says Europe should focus on data to keep AI control

In short: Mario Draghi says Europe can use AI to boost growth, but it needs more control over its own data and the computers that process it.

What's going on

Mario Draghi, former head of the European Central Bank and former prime minister of Italy, writes that Europe faces a trade-off. It wants faster economic growth, but it also wants to stay independent and protect its social and environmental goals. He argues that wider use of artificial intelligence could help Europe work more efficiently and raise productivity.

Draghi says Europe does not control much of the global AI supply chain. That includes the biggest AI labs and the most advanced computer chips. He warns this could leave Europe dependent on foreign companies and governments for key AI tools, similar to relying on someone else for access to a payment network (if access is cut, many daily activities can grind to a halt).

He says Europe’s best chance to keep some control is data. Data is the raw material AI learns from, like ingredients in a recipe. Europe has large pools of public data, like health records and statistics, and lots of industrial data from factories. The European Commission estimates Europe’s “data economy” could pass €800 billion by 2030.

To use that data while keeping it under European rules, Draghi argues Europe needs more large data centers, meaning big buildings full of computers that store and process information. He notes the EU has under 5 percent of the world’s AI computing power, while the US has about 75 percent. He also points to long permit and grid connection delays, such as data centers taking much longer to build in Germany than in the US.

What to watch

Draghi backs a plan where big European companies sign multi-year buying commitments together, so investors will fund more European-run data centers. A recent example is a group including ASML, Capgemini, and Amadeus supporting Mistral’s planned “European Compute Units,” aiming for 1 gigawatt of capacity by 2030.

Source: Financial Times