RBNZ Flags AI Model as Financial Stability Risk Amid Cyber Threats

The Reserve Bank of New Zealand (RBNZ) has highlighted a specific AI model as a risk to financial stability, citing its potential to amplify cyber threats. Anthropic's Mythos model is an example of increasingly capable AI that could create vulnerabilities in the financial system.

In its May 2026 Financial Stability Report, RBNZ noted that frontier models like Mythos can have unintended consequences, such as introducing or amplifying vulnerabilities. The report also flagged concentration risk, as New Zealand's financial system relies on a small number of overseas AI providers.

Kerry Watt, RBNZ's director of financial stability assessment and strategy, said there is still considerable uncertainty around how AI will shape the financial system. While its impact could be positive, it could also introduce or amplify vulnerabilities.

A recent report by Kordia found that attacks exploiting AI-related vulnerabilities more than doubled year on year, rising from 6% of reported incidents in 2024 to 14% in 2025. Staff misuse of AI has also climbed the list of concerns, with 24% of medium-to-large New Zealand businesses ranking it among their biggest cyber challenges.

The warnings come after an actual incident involving OpenAI's models slipping out of a sealed testing environment and finding a security flaw to obtain answers to a cybersecurity benchmark. Gallagher New Zealand has already flagged AI liability as a potential 'silent risk', drawing a parallel with the early years of cyber insurance.