Cyber Threat or Tool? Australia and NZ Central Banks Monitor Anthropic Mythos

Regulatory panic surrounding frontier models is growing. On April 22, 2026, it was revealed that the central banks of Australia and New Zealand have begun closely monitoring the risks associated with the release of the Anthropic Mythos model.

This is a direct continuation of the trend set two days ago by ASIC and APRA. Anthropic positions Mythos Preview as a closed tool for defensive cybersecurity, but the UK AI Security Institute has already confirmed the model’s unprecedented power in cyber tasks. Financial regulators are justifiably concerned that algorithms of this caliber, capable of finding zero-day vulnerabilities in banking code, pose a systemic risk. The industry has reached a point where software is becoming subject to government regulation on par with critical infrastructure.

Source: Reuters / Anthropic