Singapore’s banking sector is moving to sharpen its cyber defences as advanced AI models raise fresh concerns over how quickly software vulnerabilities could be found and exploited, Reuters reported.

Ong-Ang Ai Boon, Director of the Association of Banks in Singapore (ABS), shared in a written response that the industry body is helping banks compare threat signals and shape sector-wide safeguards.

Individual banks have also upgraded their cyber operations to shorten the window between detecting a threat and resolving it.

The concerns centre on Anthropic’s Claude Mythos Preview, a general-purpose AI model whose ability to identify vulnerabilities in complex software systems has led the company to limit access.

Anthropic announced the model in April and launched Project Glasswing, a limited-access cybersecurity programme that gives selected partners access to Claude Mythos Preview for defensive security work.

MAS has also been coordinating with the Cyber Security Agency of Singapore to support critical infrastructure operators.

The regulator also called on financial institutions to review their systems, close known gaps quickly and maintain strong cyber hygiene.

Asian financial regulators had already signalled concern the week prior, with several indicating that they were assessing the implications of Mythos and monitoring potential risks.

Featured image: Edited by Fintech News Singapore, based on image by Golden-Mock via Magnific