Claude Mythos: Anthropic to Brief G20 Central Banks

The information, disclosed by the Financial Times on May 18, underscores the magnitude of concerns now surrounding Claude Mythos among the upper echelons of global financial regulation.

Anthropic, the company behind it, is set to brief finance ministries and the central banks that are members of the Financial Stability Board (FSB) on the cybersecurity vulnerabilities detected by its “frontier model” in the global financial system.

At the origin of this initiative: Andrew Bailey, Governor of the Bank of England and chair of the FSB, the global watchdog tasked with coordinating financial regulation for the G20 countries. He is said to have asked Anthropic to come present the capabilities of its Claude Mythos Preview model before the FSB’s members.

Also read: The EU gains access to OpenAI’s cyber model

Designed specifically for cybersecurity, it can detect vulnerabilities that have persisted for decades across web browsers, infrastructure, and software.

Given this potential, the dissemination of the model is currently tightly controlled. Only around forty organizations have access, mostly American: Amazon, Microsoft, and JPMorgan Chase are among the beneficiaries, who can thus remediate the identified vulnerabilities. Anthropic has pledged not to broaden this distribution, at the White House’s request.

A Model with Ultra-Restricted Distribution

For non-American actors, the company has nonetheless agreed to provide high-level briefings to certain institutions, such as the European Commission.

Global financial regulators are particularly exposed to this risk. Banks rely largely on aging information technology systems that tools like Mythos could exploit or, conversely, help secure.

The FSB is currently working on a report outlining “best practices” for AI adoption in the financial sector, with public consultation planned for next month.

Beyond central banks, the entire architecture of international finance is at stake.

Earlier this May, the International Monetary Fund warned about a systemic threat, noting that new AI models “raise cyber risk to the level of a potential macro-financial shock.”

Also read: Oracle launches its “Patch Tuesday”: the Claude Mythos effect?

In a blog post, its economists stressed: “Cyber risk does not respect borders. Emerging and developing countries, which often have more limited resources, can be disproportionately exposed.”

The FSB and Anthropic declined to comment on their recent exchanges.

Dawn Liphardt

Dawn Liphardt

I'm Dawn Liphardt, the founder and lead writer of this publication. With a background in philosophy and a deep interest in the social impact of technology, I started this platform to explore how innovation shapes — and sometimes disrupts — the world we live in. My work focuses on critical, human-centered storytelling at the frontier of artificial intelligence and emerging tech.