**
As the world grapples with the implications of rapidly advancing artificial intelligence, the Governor of the Bank of Canada, Tiff Macklem, has urged swift action from governments and regulators regarding the cybersecurity challenges posed by Anthropic’s latest AI model, Claude Mythos. Speaking from Washington during the spring meetings of the International Monetary Fund and World Bank, Macklem highlighted the urgency of addressing these concerns alongside ongoing global issues, such as the war in Iran.
AI’s Emerging Threat Landscape
Macklem’s remarks come in the wake of Anthropic’s announcement of the Mythos model earlier this month, a powerful AI tool that has demonstrated a remarkable ability to identify and exploit software vulnerabilities. However, the company has hesitated to release it widely due to potential cybersecurity risks. “The world is moving quickly. We need to keep up. We’re going to need to move quickly,” Macklem told reporters, indicating that the financial sector must adapt to the evolving threat landscape.
This week, Macklem engaged in discussions with U.S. Federal Reserve Chair Jerome Powell concerning Mythos, while Canada’s Finance Minister François-Philippe Champagne held talks with U.S. Treasury Secretary Scott Bessent. The urgency of these discussions was underscored by the Canadian Financial Sector Resiliency Group (CFRG), which met for the second time this week to evaluate the security implications of Mythos.
The Implications of Mythos
Macklem acknowledged that, while there is currently no active cyber incident linked to Mythos, the model has the potential to expose and exploit vulnerabilities at an unprecedented pace. “We’re not responding to an immediate problem. But the potential for Mythos, and just new AI large language models in general, is that they can expose vulnerabilities much faster and exploit them much faster,” he explained. This reality necessitates that organisations enhance their cybersecurity measures urgently.
The discussions surrounding the financial system’s potential vulnerabilities intensified following a meeting convened by Powell and Bessent with leading U.S. bank CEOs. This gathering prompted central banks and regulators worldwide to reassess the nature of the cybersecurity threat posed by emerging technologies like Mythos. Macklem noted, “This isn’t a one-off. Mythos has arrived; it’s a lot more powerful than what came before.”
International Cooperation is Key
Champagne, addressing reporters, emphasised the necessity for global collaboration in tackling the cybersecurity challenges presented by swiftly evolving technologies. “This has been something that has been talked about by many in terms of the measures that we need to take, not just in North America, but in G7 countries and beyond, to maintain the integrity of our financial institutions,” he stated.
He described the emergence of Mythos as a “test case” for the agility of governments, corporations, and regulators in responding to rapid technological advancements. “I’m also mindful of the next wave, which could be when you couple AI, quantum [computing], and cyber together,” he added, highlighting the imperative for proactive strategies.
Federal AI Minister Evan Solomon met with Anthropic representatives earlier this week. He expressed support for the company’s decision to initially release a preview version of Mythos to a select group of firms, which includes major players like Amazon, Microsoft, and JPMorgan Chase. This project, known as Project Glasswing, aims to strengthen cyber defences before AI-driven attacks can exploit vulnerabilities.
A New Era of Cybersecurity Threats
Anthropic has refrained from making Mythos broadly available, citing its potential to quickly find and exploit software flaws—some of which were previously undiscovered. The AI Security Institute in the UK recently published an assessment indicating that Mythos surpasses other models in autonomously exploiting software vulnerabilities, particularly within weakly defended systems.
While researchers are cautious about Mythos’s ability to attack more robust systems, the model has already identified thousands of vulnerabilities across major operating systems and web browsers.
Why it Matters
The emergence of Anthropic’s Claude Mythos marks a significant turning point in the intersection of artificial intelligence and cybersecurity, presenting both unprecedented challenges and opportunities. As financial institutions and regulators scramble to develop strategies for mitigating these risks, the need for rapid adaptation and international cooperation becomes more critical than ever. The stakes are high: failure to address these emerging threats could undermine the integrity of global financial systems and expose sensitive data to malicious actors. As we stand on the precipice of this new technological frontier, the urgency for robust cybersecurity frameworks has never been clearer.