TAIPEI (Taiwan News) — The Financial Supervisory Commission said Tuesday that a new AI model can rapidly identify cybersecurity weaknesses, urging financial institutions to strengthen defenses.
The FSC said Anthropic’s Claude Mythos Preview, introduced in April, can rapidly identify cybersecurity weaknesses, increasing the risk of AI-driven attacks, per CNA. The regulator urged financial institutions to adopt AI-powered security tools to “fight AI with AI.”
FSC Department of Information Management Director-General Lin Yu-tai (林裕泰) explained that advanced AI models with agent capabilities can uncover vulnerabilities far more quickly than traditional methods. He said AI-powered cyberattacks are a reality and regulators worldwide are discussing how to address the threat.
The FSC called on financial institutions to accelerate implementation of its financial cybersecurity resilience blueprint by adopting zero-trust architecture, strengthening monitoring, and improving threat detection.
The regulator also urged firms to review their cybersecurity environments within three to six months, identify system exposure, and adopt measures such as system isolation and access controls.
The FSC said financial institutions should incorporate AI into their long-term cybersecurity strategies, including vulnerability detection and incident response. It added that it will continue monitoring advances in AI to strengthen Taiwan’s financial sector.
