France’s biggest bank, BNP Paribas, is intensifying efforts to harden its cyber defenses as a new generation of powerful artificial intelligence models begins reshaping the global security landscape and exposing vulnerabilities at unprecedented speed.
Marc Camus, the bank’s chief information officer, said the rise of advanced AI systems capable of rapidly identifying software weaknesses has fundamentally altered how financial institutions must think about cyber resilience. Banks are now preparing for an environment in which AI tools can uncover thousands of vulnerabilities simultaneously, compressing the time available to detect, prioritize, and patch critical flaws.
“There is a lot of noise in the market on Mythos and the fact that Mythos is accessible or not accessible for some banks, particularly European banks,” Camus said during a joint press conference on Tuesday with French AI startup Mistral AI.
The comments underline growing anxiety among European lenders that U.S. financial institutions could gain a strategic edge in cybersecurity if they obtain earlier or broader access to frontier AI systems such as Mythos, the advanced cybersecurity-focused model launched earlier this year by Anthropic.
Mythos was designed to identify software vulnerabilities at massive scale and speed, raising concerns across the banking sector that such systems could eventually become dual-use technologies: defensive tools for institutions but also potentially powerful assets for cybercriminals and state-backed attackers.
Camus said the key challenge facing banks is operational velocity.
“The game changer is the speed at which we have to address vulnerabilities and the scale. There are lots of them discovered at once,” he said.
“So we need to prepare ourselves for that and that’s something we are really working on very, very hard.”
The warning emerges through a broader transformation underway in global finance as banks increasingly integrate generative AI into core operations while simultaneously confronting the risks created by the same technology. Financial institutions are under mounting pressure to modernize legacy systems, secure sprawling digital infrastructure, and comply with tightening regulatory standards around operational resilience.
European banks have become particularly sensitive to the issue as the AI race increasingly concentrates around U.S. firms with enormous computing resources and privileged access to cutting-edge semiconductor infrastructure. Executives across the sector fear that uneven access to the most advanced AI models could widen competitiveness gaps between Wall Street banks and their European peers.
BNP Paribas’ expanding partnership with Mistral illustrates how European financial institutions are attempting to build regional AI alliances rather than relying entirely on U.S. technology giants. The French lender and Mistral said they have broadened a collaboration that began in 2023, when the Paris-based startup was still in its early stages.
Mistral has quickly emerged as one of Europe’s most prominent AI firms and is widely viewed as a strategic continental alternative to American companies such as OpenAI, Anthropic, and Google.
Corentin Petit, Mistral’s global head of solutions, said the company is tailoring its models for heavily regulated sectors, including banking, where compliance, auditability, and data governance are becoming critical differentiators.
“We will optimize on benchmarks that matter for our customers in the industry,” Petit said, adding that more details would be disclosed later.
BNP Paribas is already deploying Mistral’s technology across multiple divisions. Sophie Heller, chief transformation officer for the bank’s retail and consumer business, said the lender is using Mistral-powered tools for internal productivity systems, customer-facing virtual assistants in France and Belgium, and compliance functions at its Belgian subsidiary Fortis.
Within BNP’s investment banking division, the applications are expanding even further. Charles Holive, the unit’s chief AI officer, said the bank is using AI for document extraction, equity research support, and enterprise-wide knowledge retrieval systems serving tens of thousands of employees.
The collaboration has moved beyond a conventional vendor-client arrangement. Mistral engineers and data scientists are now embedded directly within BNP teams to jointly develop and scale AI systems internally. The partnership also comes as Europe pushes to reduce dependence on U.S. cloud and AI infrastructure. European policymakers and executives have increasingly argued that the continent risks losing technological sovereignty if its banking and industrial sectors rely too heavily on American AI providers.
Additionally, regulators are warning that AI could amplify systemic risks inside the financial sector if governance frameworks fail to keep pace. Supervisors in Europe and the United States have repeatedly cautioned banks about model hallucinations, data leakage, cyber manipulation, and concentration risk tied to a small number of dominant AI providers.
