ASIC Warns Financial Firms on Cybersecurity Threats from Frontier AI Models | 2026 – News and Statistics
Publish Date: 2026-05-08 20:31:00
Source Domain: www.indexbox.io
May 9, 2026
Australia’s corporate regulator has called on the country’s financial sector to address emerging cybersecurity threats linked to advanced AI systems like Mythos. The Australian Securities and Investments Commission (ASIC) issued a letter on Friday urging the financial services industry to strengthen its cybersecurity practices, as reported by Reuters.
ASIC Commissioner Simone Constant stated that cyber risk has entered a new phase, with frontier AI models increasing risk by exposing vulnerabilities more quickly than anticipated. The commissioner advised firms to act with discipline now rather than wait for complete clarity on the threat. Experts have warned that Mythos, an AI model with high-level coding capabilities, possesses an unprecedented ability to identify cybersecurity weaknesses.
The developer of Mythos, Anthropic, did not respond to a request for comment on ASIC’s letter. This regulatory warning follows a separate alert from Australia’s banking regulator last month, which noted that the domestic financial services industry’s information security practices were not keeping pace with changes in AI technology. Constant emphasized that the time for action is immediate, comparing the situation to a clock at ‘a minute to midnight.’
Anthropic has launched a program called Claude Mythos Preview under Project Glasswing, a tightly restricted access initiative that includes major technology firms such as Amazon, Microsoft, Nvidia, and Apple. A recent survey questioned the ability of central banks and financial regulators to monitor and counter AI-related risks, finding that authorities significantly lag financial firms in AI adoption and lack data on emerging harms. Research published in April by the Cambridge Centre for Alternative Finance showed financial institutions are adopting AI at more than double the rate of their supervisors, with only two out of ten regulators reporting advanced AI…
Source
