A new AI model called Mythos is raising alarms in the global financial world. Developed by Anthropic, this AI can spot hidden cyber flaws in software and systems used by banks and tech companies. These flaws could open doors for hackers. That’s why Anthropic is sharing its findings with the Financial Stability Board (FSB), a watchdog group made up of top finance ministries and central banks from major economies.

Mythos is not available to the public yet. Anthropic gave access only to a select group of companies, including Apple and JP Morgan. The AI scans for weaknesses in everything from web browsers to operating systems. It can find vulnerabilities that have gone unnoticed for years. Some of these flaws lie in old software that many financial institutions still depend on.

The concern is that Mythos could help hackers develop complex cyberattacks. These attacks might bypass security measures banks have relied on for decades. Experts warn this could put the entire financial system at risk. The International Monetary Fund (IMF) even called for a coordinated global response to handle these fast-moving AI risks.

Why Mythos Matters for Cybersecurity

At its core, Mythos tests the limits of what AI can do in cybersecurity. It can autonomously discover new weaknesses and create multi-step attacks that usually take human experts days to plan. The AI completed a difficult cybersecurity challenge called the “cooling tower” test. It solved this test three times out of ten, a first for any AI model.

This shows how quickly AI is advancing in cyber offense and defense. Experts say the pace of progress is measured in months now, not years. That means financial institutions must rethink their cyber defenses fast. The UK’s AI Security Institute is already developing tougher tests to keep up with these improvements.

Global Financial Leaders Take Notice

The Financial Stability Board, chaired by the Bank of England’s Governor Andrew Bailey, is taking Mythos seriously. Bailey warned that this AI could “crack the whole cyber risk world open.” He pushed Anthropic to brief the FSB on what Mythos has found. This group guides financial rules for G20 economies and helps protect the global financial system.

Leaders from JP Morgan and Goldman Sachs have spoken publicly about Mythos. JP Morgan’s CEO said AI has made cyber defense harder, but it could also help companies fight back. Goldman Sachs’ CEO said he watches Mythos’ capabilities closely. Meanwhile, UK regulators have urged financial firms to boost their basic cyber hygiene. This means patching old systems, improving authentication, and being more vigilant overall.

Cybersecurity experts caution that most breaches still come from known risks like weak passwords and outdated patches. Mythos shows there are many more hidden threats lurking in older systems. Because cyber risk crosses borders, it’s vital that countries work together to manage AI-driven threats. The IMF called inconsistent oversight a danger to the interconnected financial system.

Anthropic is working with regulators in the US, UK, and Europe to share insights responsibly. The White House has limited Mythos’ broader distribution to prevent misuse. Anthropic also plans to brief the European Commission. This controlled sharing aims to help financial institutions prepare without letting hackers access the model.

In short, Mythos is a powerful new tool that uncovers cyber risks no one saw before. It’s forcing financial watchdogs and banks to rethink how they protect their systems. The AI revolution in cybersecurity is underway, and Mythos is at the frontline of this change.