World Federation of Exchanges publishes Mythos AI paper and urges stronger existing cyber resilience frameworks

The World Federation of Exchanges has published a position paper on Anthropic's Mythos model that sets out an initial industry view on recent AI and cybersecurity developments. The paper says announcements about Mythos warrant serious attention, but based on current information the most effective near-term response for financial market infrastructures is to keep strengthening existing cyber resilience, operational resilience and vulnerability management frameworks rather than treat the model as a fundamentally new category of threat. Drawing on analysis by the WFE Cybersecurity Working Group, the paper says AI models are increasingly able to identify vulnerabilities and automate known attack techniques, expanding their potential role in cyber operations. It points market infrastructures toward rigorous vulnerability management, patching, identity and access controls, software supply-chain assurance, and incident detection and response. It also highlights the potential for such tools to help cybersecurity professionals identify vulnerabilities, while calling for continued industry coordination and information-sharing to track how the threat landscape evolves. As next steps, the WFE said it will continue to assess emerging cybersecurity issues through its Cybersecurity Working Group, including gathering intelligence from members on cyber incidents and attacker techniques. The group will also continue to feed practitioner input into regulatory developments to inform future policy discussions.