Advanced AI Bank Hacker Leaked, Causing The Fed Boss to Panic

Jakarta, CNBC Indonesia - Mythos, an AI programme feared by banks worldwide, has been leaked. Several unidentified users are said to have successfully accessed the AI model created by Anthropic.

Bloomberg News, quoted by Reuters, reported that several accounts on online forums managed to access Mythos on the same day Anthropic announced a limited release for testing.

Since that day, Mythos has been continuously used by those unidentified users. However, its usage is not related to cybersecurity.

“We are investigating reports claiming unauthorised access to Claude Mythos Preview via third parties,” said an Anthropic spokesperson.

Mythos is part of Project Glasswing, Anthropic’s ambition to build an AI defence system involving several institutions to test Claude Mythos Preview.

The emergence of Mythos has made regulators around the world anxious because the AI model has the capability to quickly identify holes in digital security systems.

A previous Reuters report also stated that US Treasury Secretary Scott Bessent and Federal Reserve Chair Jerome Powell met with the bosses of major banks to discuss the potential risks from Mythos.

Not only in the US, concerns about Mythos have also reached the UK. Local authorities have spoken with major banks and cybersecurity officials about potential risks that may arise.

Security measures are also being taken in Germany. The President of the German Banking Association and CEO of Deutsche Bank, Christian Sewing, said banks have been in close contact with European regulators regarding Mythos.