France's government will hire domestic AI providers such as Mistral for its internal tools, Budget Minister David Amiel said Tuesday. He explicitly ruled out using OpenAI, the maker of ChatGPT.
"This excludes OpenAI," Amiel said.
The announcement follows a cyberattack on the French tax agency that exposed data belonging to roughly 700,000 taxpayers. The Finance Ministry disclosed the breach Thursday, and Amiel said the government plans to use AI to identify vulnerabilities in its services against future attacks.
Amelie Verdier, head of the French tax agency, said her teams detected another data breach on Monday. That attack is still being evaluated, she added.
What the government said
Amiel's remarks came after the weekly cabinet meeting at the Elysee Palace. He framed the move as a matter of sovereignty, preferring French providers for government work. The choice also signals a preference for European AI models over American ones in sensitive public-sector applications.
The decision affects how France's public administration will handle AI-assisted security and internal operations. Government employees working with taxpayer data or public records should expect Mistral-based tools in their workflow rather than OpenAI products.
What happens next
The tax agency is still assessing the scope of the second breach detected Monday. The first attack compromised personal data of approximately 700,000 taxpayers, according to the Finance Ministry.
For those working in government, the practical takeaway is straightforward: French institutions are moving toward domestic AI vendors for official use, and cyber defense is the immediate priority. That means training efforts for public-sector staff will likely center on Mistral's tools and on security practices tied to AI-assisted threat detection. Public employees handling sensitive data should prepare for stricter AI usage policies and vendor-specific workflows in the coming months.
For AI for Government contexts, this shift also underscores the importance of understanding which AI providers are approved for official work. And given the cyberattack response, AI for Cybersecurity Analysts training will likely become more relevant for French public-sector IT teams.
Your membership also unlocks:
