CSA opens public consultation for additional guidelines on agentic AI
Feedback aims to improve AI development and fraud detection systems, amongst others.
The Cyber Security Agency of Singapore (CSA) opened for public consultation its guidelines of artificial intelligence (AI) security as the use of agentic AI rises.
The addendum outlines how risks can be identified and assessed and provides controls to mitigate the risks, which may be applied to app development, client onboarding systems, and automated fraud detection systems.
The addendum on agentic AI systems was developed with industry and government, and it is designed to be read alongside agentic AI guidelines, the CSA said.
Public consultation on the addendum will run from 22 October to 31 December. Feedback is open to the AI and cybersecurity community, and this can be submitted to CSA’s email.
The consultation was opened at the Singapore International Cyber Week 2025.