Introduction#
Switzerland's financial regulator, Finma, has raised alarms about the potential risks associated with Anthropic PBC's artificial intelligence tool, Mythos. The regulator believes that granting banks quick access to this powerful AI could jeopardize the stability of the country's financial system.
Systemic Risk Concerns#
A spokesperson from Finma stated that the immediate and uncontrolled availability of AI models like Mythos could be classified as a systemic risk. This means that if such technology were to be misused, it could expose existing software systems to numerous unknown vulnerabilities, which could be exploited rapidly using AI. Essentially, this could lead to widespread issues across various financial institutions simultaneously.
Limited Access to Mythos#
Anthropic has indicated that Mythos is too potent to be made available to the general public. Instead, the company plans to restrict access to a select group of carefully chosen parties. They emphasize that while Mythos is effective at identifying software vulnerabilities, if it were to fall into the wrong hands, it could empower attackers to steal sensitive data or disrupt critical infrastructure.
Regulatory Measures and Industry Response#
In light of these concerns, regulators, central bankers, and corporate leaders are seeking a deeper understanding of AI technology. There is a growing worry that financial systems outside the U.S., particularly in Europe, may be at a disadvantage due to limited access to such advanced tools. Bundesbank President Joachim Nagel highlighted the need to prevent the misuse of AI while ensuring that relevant institutions have access to avoid competitive imbalances.
Finma is actively engaging with the Federal Office for Cybersecurity, banks, and critical service providers to address these challenges. They stress the importance of incorporating the evolving threat landscape into risk management strategies, as cyber attacks are becoming increasingly sophisticated and easier to execute with AI assistance.