The financial services industry faces an urgent technological arms race. Marlene Amstad, president of Switzerland's market regulator FINMA, has warned that banks and financial supervisors must accelerate their adoption of advanced artificial intelligence systems to counteract rapidly evolving cybersecurity threats. Speaking following an international hackathon convened to develop new regulatory technology, Amstad underscored the critical imperative facing the global banking sector as artificial intelligence simultaneously creates new vulnerabilities whilst offering defensive solutions.
The pressure on financial institutions stems from a fundamental asymmetry in cybersecurity dynamics. As malicious actors harness AI capabilities to identify and exploit system weaknesses with increasing sophistication, traditional defensive measures have become inadequate. Amstad emphasised that "as hackers move faster, banks must adapt by patching vulnerabilities more rapidly." This acceleration requirement extends beyond individual institutions to encompass entire regulatory frameworks, forcing supervisors worldwide to reconceptualise their oversight approaches and deploy tools that can match the pace of technologically sophisticated attackers.
FINMA has taken a leading role in coordinating this global response by helping establish a dedicated forum within the International Organization of Securities Commissions, the preeminent standard-setter for financial market regulation. This initiative reflects the recognition that cybersecurity challenges transcend national boundaries and require coordinated international solutions. The forum brings together supervisory authorities that collectively oversee approximately 95 percent of global financial markets, creating an unprecedented platform for sharing threat intelligence, developing common standards, and jointly building protective technologies.
The scope of the regulatory response became apparent during this week's hackathon, which convened approximately 100 policy specialists and technology experts with an explicit mandate to design and prototype new supervision tools. The initial focus centred on cryptocurrency markets, an area where regulatory technology remains comparatively nascent and where the combination of AI-driven risks and existing supervisory gaps presents particular challenges. By bringing regulators, technologists, and policy experts into collaborative development sessions, authorities aim to accelerate the translation of technological innovation into practical supervisory capability.
Beyond identifying vulnerabilities, regulators are exploring more fundamental architectural solutions. Amstad indicated that watchdogs are investigating the possibility of embedding safeguards directly into digital asset systems themselves, rather than relying solely on external monitoring. This proactive approach represents a conceptual shift from reactive oversight toward preventive system design, acknowledging that some cybersecurity challenges may be more effectively addressed through technology configuration rather than human oversight alone.
The risks posed by advanced AI systems themselves have become apparent through recent operational experience. Amstad highlighted how exposure to sophisticated models such as Anthropic's Mythos has revealed previously hidden vulnerabilities and operational risks within financial institutions. These discoveries have exposed the paradoxical nature of AI in financial supervision: the very systems designed to strengthen defences can themselves become vectors for understanding and exploiting weaknesses. This dual-use characteristic requires regulators to balance access to cutting-edge technology against the security implications of deploying such systems.
Geopolitical tensions surrounding advanced AI access add another layer of complexity to the regulatory landscape. The U.S. government recently ordered Anthropic to suspend exports of its Mythos and Fable AI models, citing national security considerations. This action illustrates how financial sector technology policy has become entangled with broader strategic competition over artificial intelligence capabilities. Such trade restrictions risk fragmenting the global financial regulatory framework and creating asymmetries in supervisory capability across different jurisdictions.
China's emergence as a developing source of advanced AI systems underscores the shifting technological landscape. Chinese cybersecurity firm 360 Security Technology announced this week that it has developed a domestic equivalent to Mythos, suggesting that geopolitical constraints on technology access may accelerate the development of alternative technological ecosystems. For regulators in smaller economies and developing markets, these trends raise questions about technology sovereignty and the capacity to maintain independence from larger powers controlling access to critical AI systems.
Amstad has articulated a compelling case for why Switzerland and other financial jurisdictions must retain access to the most advanced AI models available globally. She contends that regulatory effectiveness increasingly depends on deploying state-of-the-art technology to harden financial systems before they encounter real-world threats. This argument positions technological access not merely as a competitive advantage but as a fundamental requirement for maintaining financial stability and protecting citizens' assets. The implicit warning is that jurisdictions denied access to advanced AI tools risk falling behind in an ongoing competition for regulatory excellence.
For Southeast Asian countries and other emerging markets, the implications of this global regulatory transformation deserve careful consideration. Many regional financial authorities lack the technological resources and expertise concentrated in Switzerland, Singapore, and other advanced jurisdictions. As international standards increasingly presume access to sophisticated AI-based supervision, emerging market regulators face mounting pressure to upgrade their technological capabilities or risk isolation from global financial governance frameworks. The potential for creating a two-tier regulatory system—where wealthy jurisdictions possess advanced AI tools whilst developing nations rely on legacy systems—represents a significant governance risk for global financial stability.
The collaborative approach embedded in FINMA's international forum offers a potential path forward for more equitable technological distribution. By developing tools through multilateral cooperation and sharing knowledge across the 95 percent of markets represented, supervisors can potentially democratise access to critical regulatory technology. However, the success of such initiatives depends on sustained political commitment and the willingness of technology leaders to prioritise global financial stability over narrow competitive or geopolitical interests.
The financial sector's rush to deploy AI-based supervisory tools reflects broader recognition that the traditional regulatory paradigm has become obsolete. As artificial intelligence capabilities advance exponentially, regulatory frameworks designed for earlier technological eras cannot adequately protect modern financial systems. The challenge ahead involves not merely adopting new tools but fundamentally rethinking supervisory philosophy, governance structures, and international coordination mechanisms to ensure that regulatory capacity can evolve at least as rapidly as the technologies and threats it must manage.
