Australia Partners with Anthropic to Anticipate Cyber Security Gaps from AI Mythos

Australian Banking Association CEO Simon Birmingham (photo: x @strangerous10)

JAKARTA - The Australian government has begun partnering with technology companies including Anthropic to anticipate potential cyber security vulnerabilities, following the limited launch of the Claude Mythos artificial intelligence model which triggered global concerns.

The move was confirmed by a spokesperson for the Minister of Home Affairs, Tony Burke, on Thursday, April 23, who confirmed that the government was working with software providers to map the new risks of advanced generative AI technology.

The Claude Mythos model, which was released in a limited preview version, is designed to help with cyber defense tasks. However, its capabilities actually triggered an alarm after the company claimed that the system was able to find "thousands" of major security holes in various major operating systems and browsers.

"Protecting critical infrastructure is a top priority, and we ensure that any potential vulnerabilities can be identified early," said a government representative.

Mythos development is carried out through a limited program called Project Glasswing, which involves a number of global technology giants such as Amazon, Microsoft, Nvidia, and Apple Inc. In addition, access is also expanded to more than 40 organizations that play a role in the development and maintenance of critical software infrastructure.

Central banks in Australia and New Zealand are also monitoring this development and coordinating with global regulators. The main concern comes from the potential use of AI to accelerate cyber attacks, especially in sectors with complex systems such as banking.

Australian Banking Association CEO Simon Birmingham said the banking industry was now actively coordinating with regulators to ensure the stability of the financial system remained intact.

Experts assess that AI capabilities such as Mythos in automatically analyzing code can be a "double-edged sword". On the one hand, this technology accelerates the identification of security gaps. But on the other hand, if misused, AI can accelerate the development of more sophisticated and difficult-to-detect cyber attacks.

This phenomenon marks a new phase in the global cyber security race, where AI is not only a defensive tool, but also a potential threat. Governments and industries are now racing to ensure that the technology remains on the right side.

Follow VOI Whatsapp Channel