ZURICH — Financial institutions and their regulators must act swiftly to embrace new technology and close system vulnerabilities, as artificial intelligence continues to amplify cybersecurity threats across the globe. That warning came from a top Swiss financial watchdog in a recent interview.
Marlene Amstad, who serves as president of Swiss market regulator FINMA and chairs an international forum focused on supervisory technology, spoke with Reuters after an initial hackathon event designed to help market supervisors build new oversight tools.
Software vulnerability detection models have recently flagged a sharp rise in cyberattack and national security risks, with AI raising serious questions about safety and accountability within financial institutions.
“As hackers move faster, banks must adapt by patching vulnerabilities more rapidly,” Amstad said.
FINMA played a key role in establishing a forum within the International Organization of Securities Commissions — a global standard-setting body for market regulation — with the goal of encouraging watchdogs covering approximately 95% of worldwide financial markets to adopt AI tools.
Roughly 100 policy and technology specialists gathered this week for the hackathon, working together to develop tools specifically aimed at supervising cryptocurrency markets, Amstad noted. She added that regulators are also exploring the possibility of embedding safety measures directly into digital asset systems.
Amstad pointed to experience with models such as Anthropic’s Mythos as having revealed AI-related operational risks by exposing system vulnerabilities. This comes as the U.S. government this month ordered Anthropic to halt exports of its latest Mythos and Fable AI models, citing national security concerns.
Meanwhile, Chinese cybersecurity company 360 Security Technology announced this week that it has built a domestic alternative to Mythos.
“Switzerland must retain access to the most advanced AI models,” Amstad said, adding that AI will play a critical role in strengthening systems before they are put into use.
