Global finance chiefs outside the United States are expressing significant concern over potential unequal access to information regarding Anthropic’s unreleased Mythos artificial intelligence (AI) model. A Bloomberg report on Thursday (April 16) highlighted these anxieties, pointing to a perceived disparity in information sharing compared to their American counterparts. The level of detail Anthropic is currently sharing about the advanced AI model with entities outside the U.S. remains largely unknown, fueling calls for greater transparency.
International Calls for Transparency
Specific anxieties and calls for action have emerged from prominent non-U.S. officials. Canada’s Finance Minister Francois-Philippe Champagne stated on Wednesday (April 15) that he intends to raise the Mythos issue with his counterparts, emphasizing a critical shared objective: “We have a common interest to ensure the resiliency of our financial system.” This sentiment was echoed in discussions among Group of Seven (G7) finance chiefs at a Wednesday meeting, where officials deliberated on the necessity of an international institutional framework for the governance of AI, though concrete next steps are yet to be defined.
Several European officials have reportedly been encouraging their U.S. peers to share more information. European Central Bank President Christine Lagarde, speaking to Bloomberg Television on Tuesday (April 14) about Mythos, issued a stark warning: “If it falls in the wrong hands, it could be really bad.” Further underscoring the urgency, Sweden’s Finance Minister Elisabeth Svantesson confirmed that AI would be a key topic at an “early-warning meeting” scheduled for Thursday, bringing together central bankers and ministers.
Anthropic’s Project Glasswing Initiative
In response to the growing scrutiny surrounding powerful AI models, Anthropic unveiled a program called Project Glasswing on April 7. This initiative is designed to grant select partners early access to “Claude Mythos Preview,” enabling them to proactively identify vulnerabilities and strengthen their systems against potential threats before the model’s full release. Anthropic has also been in discussions with U.S. government officials regarding the AI model and its cyber capabilities, according to the company.
In its announcement, Anthropic articulated its broader vision for the program, stating: “We are hopeful the Project Glasswing can seed a larger effort across industry and the private sector, with all parties helping to address the biggest questions around the impact of powerful models on security.” This statement highlights the company’s recognition of the significant security implications associated with advanced AI.
Expanding Access, Persistent Concerns
Recent developments indicate an expansion of Project Glasswing. It was reported on Thursday that Anthropic is now ready to offer Mythos to British banks, allowing them to test the AI model ahead of its official release. While this move signifies a step towards broader engagement, the underlying concerns from the international financial community persist. European officials, in particular, continue to press their U.S. counterparts for more comprehensive and equitable information sharing, linking it directly to the ongoing G7 discussions about establishing a global governance framework for AI.
The tension between the proprietary development of powerful AI models like Mythos by companies such as Anthropic and the global financial system’s imperative for transparency and resilience remains a central issue. The concerted calls from international finance chiefs underscore a growing recognition of AI’s systemic importance and the urgent need for coordinated international oversight to mitigate potential risks and ensure stability across global markets.
