Stay informed with free updates
Simply sign up to the Artificial intelligence myFT Digest — delivered directly to your inbox.
The EU’s tech chief has warned that AI has become a geopolitical weapon, pushing for Europe to develop its own alternatives to US models faster or risk being cut off from strategic capabilities.
Brussels feared being “dependent on third countries for these very critical technologies”, Henna Virkkunen told the FT in an interview. Governments with the power to cut off other countries from accessing AI models could use that in their own interests, she warned.
The EU’s fears about reliance on US tech were realised in June when Washington placed export controls on Anthropic’s leading models over security concerns. The measures were lifted after a backlash from foreign governments and Silicon Valley, but the episode renewed fears that the US could turn off access to tech that underpins the European economy via so-called kill switches.
Those that controlled critical technologies were ‘‘not only ruling the economy’’ but also had ‘‘a big strategic asset’’ at the geopolitical level, Virkkunen said.
The Anthropic cut-off “very clearly tells us how important it is now in our cyber security environment to be prepared for that kind of reality”, said Virkkunen, a Finnish politician who oversees the European Commission’s tech and cyber security issues.
‘‘We know that there will be more of that kind of very capable AI models coming to the markets in the coming months and years. So we have to be very well prepared for that.”
She compared the situation with AI to a decision by the previous US administration to impose export controls on cutting-edge chips that divided EU member states into those that were given access and those that weren’t.
AI capabilities ‘‘are really strategic assets nowadays and that’s why it’s so important for us to build our own technological sovereignty”, she said.
Brussels last month presented a tech sovereignty package to loosen its dependence on US technology by backing European alternatives in sectors from semiconductors and cloud computing to AI.
The plan has incentives to accelerate the construction of European data centres and favour homegrown cloud and AI technologies, such as the AI company Mistral or cloud providers such as Scaleway or OVHcloud.
“It’s so important also that Europe is building up our own capacities and that we are not dependent on third countries for these very critical technologies,” Virkkunen said.
In order to finance those investments, the EU and the European Investment Bank will set up a new mechanism to make strategic investments in European tech companies.
Meanwhile, Brussels is pursuing bilateral talks with the US administration to secure continued access to the most advanced AI models.
The leaders of the EU, France, Germany and Italy also used a G7 summit in June to discuss with US President Donald Trump the possibility of setting up a “trusted partner” scheme, in a bid to ensure ongoing access to the most powerful AI models that have the power to discover critical cyber security gaps.
“We continue to work internationally on that [proposal],” Virkkunen said.
