Anthropic’s new AI model Claude Mythos identifies flaws in every major OS

Global financial leaders have issued a dire warning regarding Anthropic Claude Mythos Preview, a newly developed AI model capable of identifying vulnerabilities in every major operating system and web browser.

Senior officials from the IMF, World Bank, and the Bank of England expressed "unprecedented" concern during recent international summits, arguing that the model’s capabilities could severely disrupt the global banking system by exposing critical weaknesses in lenders’ cyber defences.

Andrew Bailey, Governor of the Bank of England and chair of the Financial Stability Board, called the development a "very serious challenge," noting that such powerful AI makes it dangerously easy to detect IT vulnerabilities.

European Central Bank President Christine Lagarde echoed these fears, warning that while Anthropic is acting responsibly, the model could be catastrophic if exploited by rogue actors.

Canadian Finance Minister François-Philippe Champagne took a bleaker view, comparing the AI to an "unknown threat" even more complex than current geopolitical conflicts, urging finance ministers to prioritise the resiliency of the financial system.

Recognising the severe risks to public safety and national security, Anthropic has decided against a public release for Mythos.

Instead, under Project Glasswing, the company will limit access to a select group of 40 organisations, including tech giants Google and Microsoft.

However, Anthropic issued its own stark caution, stating that these advanced cyber capabilities will inevitably spread to groups lacking a commitment to safety.

As James Wise of Balderton Capital noted, Mythos is likely "the first of what will be many more powerful models" to come.