The attempt to global harmony ended with a harsh attitude. As Emmanuel Macron’s AI summit ended on February 11, the U.S. vice president frankly proposed the U.S. vision of the most artificial intelligence, overruled Europe and in the usual Group photos left Europe before. As far as EU countries are concerned, EU countries have collaborated with China and the Global South, while highlighting the need to limit the risks of using AI.
Both Europe and the United States should reconsider their approach. After working at DeepSeek, Europe has gained unexpected opportunities to catch up with it if it could throw away its regulatory tights after working at the Chinese popular model maker. The United States can no longer show a monopoly on artificial intelligence. It should change its power over its allies.
View full image
The pace of innovation is surprising. Just six months ago, it seemed that AI needed a technological breakthrough to become widely affordable. Since then, reasoning and efficiency technologies have emerged, even if it fails to use cutting-edge American chips, allowing DeepSeek to develop models close to the border. DeepSeek is just an exhibition. Researchers everywhere are competing to improve AI efficiency. For example, people from Stanford and Washington trained models more cheaply. Once you worry, the world doesn’t have enough data to train advanced systems. Now, the use of synthetic data seems to be producing impressive results.
This is a once-in-a-lifetime opportunity for Europe, which is hopeless in AI. Contrary to Google’s search engine, network effects mean that the winner will get all, with no calculation or economics laws preventing European companies from catching up. Therefore, closing the gap is a policy issue. Mr. Macron correctly encourages investment in data centers. But it is equally important to cut the traditional tape festival to stop companies from innovating and adopting AI. The EU’s AI bill is very strict: a startup that provides AI coaching services must establish a risk management system through an account, conduct impact assessments and checks, except for skipping other baskets.
Another obstacle is privacy rules. Even big tech companies, who have a large compliance team, are now delaying the launch of their AI products in Europe. Imagine the cost of a startup. German manufacturers sit on a large amount of proprietary data that can help create productivity-enhancing AI tools. But fear of foul regulations disappointed them. Smart relaxation of rules and coordinated enforcement will help Europe leverage the potential of AI.
The United States needs to wake up, too. China’s progress shows that Uncle Sam has only held a lower monopoly power over the cutting-edge chips. Instead, it needs to attract the best talent in the world, which can be annoying for the Margo Republicans.
The United States should also change the way it interacts with its allies. In Paris, Mr. Vance correctly warned against using Chinese infrastructure (and the fact that China signed the summit’s declaration on AI governance may explain why the United States refused to do so). But if the United States prefers to let his friends use its technology, the United States will be more successful in blocking the adoption of Chinese artificial intelligence. In the last few days of the office, Joe Biden proposed strict AI controls that would hinder even allies like India’s exports. Modifications to these people will encourage countries to use American technology rather than pushing them into China’s embrace. American AI is now facing competition. If ruled supreme, Uncle Sam would have to attract rather than threaten.
