AI Is Making Banks More Vulnerable to Cyberattacks, Swiss Financial Regulator Warns

Artificial intelligence is rapidly transforming the financial sector, but it is also creating new cybersecurity risks that banks and regulators may not be equipped to handle.

That's the warning from Marlene Amstad, president of Switzerland's financial regulator FINMA, who says banks and financial watchdogs need to adopt AI-powered security tools much faster as hackers increasingly use artificial intelligence to identify and exploit vulnerabilities.

AI Is Changing the Cybersecurity Battlefield

According to Amstad, cybercriminals are now able to discover software flaws and launch attacks much more quickly with the help of AI. "As hackers move faster, banks must adapt by patching vulnerabilities more rapidly," she told Reuters, adding that financial institutions can no longer rely on traditional cybersecurity approaches alone.

The warning comes as banks worldwide continue integrating generative AI into customer service, fraud detection, lending, and internal operations, dramatically expanding their digital attack surface.

Regulators Want AI to Fight AI

Rather than viewing AI solely as a threat, regulators are increasingly looking to use the technology as part of their defence strategy.

FINMA helped establish a new supervisory technology forum within the International Organization of Securities Commissions (IOSCO), whose members oversee roughly 95% of global financial markets. The initiative is designed to encourage regulators to develop and deploy AI-powered supervisory tools.

Earlier this week, around 100 policymakers, regulators, and technology experts gathered for a hackathon focused on building AI tools for supervising cryptocurrency markets and detecting emerging financial risks.

Crypto Could Get Built-In Safety Checks

One area receiving particular attention is digital assets. According to Amstad, regulators are exploring ways to embed safeguards directly into cryptocurrency systems, allowing potential risks to be identified automatically rather than relying solely on external oversight.

The move reflects a broader shift in financial regulation, where technology is becoming part of the regulatory framework itself instead of simply being monitored from the outside.

Advanced AI Models Bring New Risks

Amstad also pointed to recent experiences with advanced AI models as evidence that the technology introduces fresh operational and security challenges.

The comments come shortly after the US government ordered Anthropic to suspend exports of its latest Mythos and Fable AI models over national security concerns. Meanwhile, Chinese cybersecurity company 360 Security Technology has announced its own domestic alternative to those models.

According to Amstad, Switzerland must continue to have access to the world's most advanced AI systems because they will play a critical role in strengthening financial infrastructure before vulnerabilities can be exploited.

Why This Matters

Banks have spent years investing in digital transformation, but the rise of generative AI is changing the nature of cyber threats faster than many organisations can respond.

AI can now automate vulnerability discovery, generate sophisticated phishing attacks, and accelerate malware development. The same technology, however, can also help detect attacks earlier, automate security updates, and identify weaknesses before criminals exploit them.

For regulators, the message is becoming increasingly clear: supervising AI may no longer be enough. They may also need to use AI themselves if they want to keep pace with the threats it creates.