SYDNEY, April 22 (Reuters) - The central banks of Australia and New Zealand said on Wednesday they were monitoring the release of Anthropic's advanced Mythos artificial intelligence model, joining authorities around the world in expressing concerns about the new cybersecurity risks it poses.
Designed for defensive cybersecurity tasks, Mythos' vast capabilities have sparked fears about the threat to traditional software security, after Anthropic said a preview had uncovered "thousands" of major vulnerabilities in "every major operating system and web browser."
Experts have also warned that the model can identify and exploit previously unknown vulnerabilities faster than companies can fix them.
The Reserve Bank of Australia said in a statement it was closely monitoring the development and was "engaging with peer regulators, government and regulated entities."
The Reserve Bank of New Zealand said it was also in contact with other regulators both domestically and in Australia over what it called the "developing risk" from Mythos.
On Tuesday, Bundesbank President Joachim Nagel called the model a double-edged sword, saying: "it could be used not only to improve digital security systems, but also to leverage their vulnerabilities for malicious purposes."
Anthropic has introduced Claude Mythos Preview through a tightly controlled program called Project Glasswing. Access has been granted to major technology companies including Amazon, Microsoft, Nvidia, and Apple.
The company has also expanded access to more than 40 additional organisations that build or maintain critical software infrastructure.
Experts say Mythos' advanced coding and autonomous capabilities could significantly accelerate sophisticated cyberattacks, especially in sectors like banking, where complex, interconnected, and often decades-old systems remain common.
(Reporting by Stella Qiu in Sydney; Writing by Alasdair Pal; Editing by Edwina Gibbs)
