The idea of European digital sovereignty, and particularly in the field of artificial intelligence (AI), is slow to materialize. However, it is decisive in terms of economic independence, industrial competitiveness and security. The legislative plan presented at the beginning of June by the European Commission already proposed, in addition to accelerating investments in precious semiconductors, to reserve sensitive contracts concerning AI and the cloud – remote computing – for European suppliers. It then revealed the extra-European dependence of the Twenty-Seven at “more than 80% for its digital products, services and infrastructures”.
The objective of independence, but also of diversification of suppliers, could find a boost in the “Anthropic affair”. The American start-up suspended access to two of its most powerful AI models, after Donald Trump’s government asked it to exclude foreign nationals. With both the worst and the best nestling in these technological advances, these new models supposed to detect cybersecurity vulnerabilities could well also constitute weapons for large-scale hacking.
The reason given, national security, is also one of the reasons why the French government has just announced additional funding for the development of European models. But also the break with the American data analysis giant Palantir, used in French intelligence. Contributing to further lifting the taboo, while refuting it, of European preference and protectionism. Faced with the American and Chinese behemoths, the French strategy must in any case imperatively be part of a strong and proactive European digital policy, which does not close the door to regulation.
