The French Center for AI Safety (CeSIA) publishes the results of a survey on how the French public perceives the risks and governance of artificial intelligence, conducted in partnership with the polling firm OpinionWay. Based on a nationally representative sample of 2,065 adults, this is the first detailed study of French citizens’ views on AI-related risks, the pace of development, regulatory measures, and the role they expect government to play in steering the technology.
1. The French public is deeply concerned about AI risks
French citizens are acutely aware of the dangers associated with AI development. Their concerns extend beyond the harms already visible today—such as deepfakes, algorithmic bias and copyright infringement—to threats that have not yet fully materialised but could emerge in the coming months or years absent adequate regulation.
The risks that worry the French most include disinformation and opinion manipulation (82%), cyberattacks (80%), the use of AI to help design weapons (78%), autonomous weapons and military escalation (75%), and loss of control over AI systems (75%).
2. A massive, cross-partisan demand for regulation
While French voters are increasingly divided on many issues, the rapid development of artificial intelligence unites them—against it. Only 8% of respondents want to accelerate AI development, 24% would keep the current pace, and 42% would rather slow it down significantly or pause it altogether.
3. The French want government—not industry—to steer AI
Regardless of political affiliation, a large majority of French citizens would rather see government, not private companies, set the direction for AI development (56% vs. 25%).
78% of respondents support international agreements to ban AI uses that threaten human life or fundamental rights (only 9% are opposed). A similar share would favour stronger corporate liability for AI-related harm and mandatory independent safety audits for the most advanced AI systems.
4. A growing gap between public opinion and political action
These findings reveal a striking disconnect between what citizens want and what their leaders are doing. While only 8% of French respondents want to speed up AI development, European leaders have made acceleration their watchword. “Europe will accelerate, France will accelerate,” declared President Macron at the AI Action Summit in February 2025, as European Commission President von der Leyen pledged to mobilise €200 billion in AI investment—including €50 billion in public funds—with no mention of safety or ethics conditionality.
“The French are remarkably united against the rapid and unchecked development of artificial intelligence. An overwhelming majority favour ethics and safety over the speed of progress. This preference cuts across all traditional political divides.”
While the French public overwhelmingly favours stronger oversight and a greater role for government in steering AI, the political momentum is heading in the opposite direction. The “Digital Omnibus” regulation proposed by the European Commission, backed by France and nearing adoption, delays by nearly two years the enforcement of the EU AI Act’s obligations for high-risk AI systems. These decisions are sharply at odds with the preferences expressed by a large majority of French citizens, who would rather see AI development slowed down and more tightly regulated.