Technology · Singapore Bureau
Singapore urges strict guardrails on autonomous AI agent powers
Authorities in Singapore have called for robust limits on the autonomy and account access granted to personal AI agents, citing emerging risks from the technology. The push comes as AI agents increasingly handle sensitive tasks like email management and financial transactions on behalf of users.
LSN Singapore ·
Singapore's regulatory authorities have issued guidance recommending tight controls over personal AI agents that perform actions autonomously on users' behalf, including website browsing, email management and booking services.
The advisory reflects growing concerns about the scope of access these systems receive to personal accounts and digital services. As AI agents become more sophisticated and integrated into daily digital life, regulators worry that insufficient safeguards could expose users to fraud, unauthorised transactions and privacy breaches.
The recommendations focus on establishing clear boundaries around agent autonomy, requiring explicit user authorisation for sensitive actions, and implementing transparent logging of all agent activities. Authorities have also emphasised the importance of robust authentication mechanisms and the ability for users to revoke agent access promptly.
The guidance aligns with Singapore's broader approach to AI governance, balancing innovation with consumer protection. Industry stakeholders are expected to factor the recommendations into product design and deployment strategies.
Experts suggest the move signals regulators' intent to establish a framework for autonomous AI before widespread adoption creates compliance challenges. Companies developing personal AI agents are being urged to engage proactively with regulatory guidance to ensure their systems meet emerging safety standards.