Singapore authorities issued a joint advisory on November 6 warning consumers that personal artificial intelligence agents pose distinct security and privacy risks when granted broad access to online accounts and applications. Unlike conventional chatbots that merely answer questions, tools such as Meta Muse, Dots, Hermes, Grok Bot, and OpenClaw can browse websites, manage emails, make bookings, and execute transactions on a user’s behalf.
Security Precautions for Personal AI Assistants
The advisory, published by the Cyber Security Agency of Singapore (CSA), the Government Technology Agency (GovTech), and the Infocomm Media Development Authority (IMDA), cautions that these agents can operate in the background and may be manipulated by malicious actors through hidden instructions on websites or in emails.
Depending on the access you give them, AI agents may read and reveal sensitive personal information, make mistakes, or take actions that you did not intend,
Cyber Security Agency of Singapore, GovTech, and Infocomm Media Development Authority
To mitigate these vulnerabilities, officials outlined five key precautionary steps. Users should connect only necessary applications, avoid granting access to banking, government, or healthcare accounts, and protect credentials such as passwords and one-time passwords. Consumers were warned that they remain accountable for any actions executed by their agents.
You remain responsible for how you use AI agents, including the actions these agents take on your behalf. Depending on the circumstances and applicable laws, you may also be held liable for these actions.
Cyber Security Agency of Singapore, GovTech, and Infocomm Media Development Authority
Singapore Builds GovHarness Platform to Secure Public Sector AI
While consumers face personal security decisions, the Singaporean government is simultaneously developing its own infrastructure to manage agentic AI use within public agencies. Speaking at the Government Technology Agency’s STACK technology conference, Senior Minister of State for Digital Development and Information Tan Kiat How announced that authorities are building a shared technology platform named GovHarness. The system incorporates common safeguards and keeps detailed logs of actions taken by public sector AI tools to ensure anomalies can be detected promptly.
We must therefore be clear about which actions an agent is authorised to take, which tools it may use and when a person must approve its actions.
Tan Kiat How, Senior Minister of State for Digital Development and Information
GovTech has also formulated an Agentic Risk and Capability Framework to govern digital assistants like Desk, establishing stricter human approval controls for high-risk operations such as financial transactions and official communications. These regulatory steps build upon the Infocomm Media Development Authority’s Model AI Governance Framework for Agentic AI, which was launched earlier in the year to assist businesses with risk management.
Parliamentary Discussions on National Digital Identity
In Parliament, lawmakers addressed broader strategies for safe adoption, responding to proposals from Marine Parade-Braddell Heights MP Tin Pei Ling regarding a national digital identity registry for AI agents. Senior Minister of State for Digital Development and Information Jasmin Lau stated that the government is collaborating with industry partners and regulators to build trust and standardise agent verification.
Private and public sector initiatives are already underway. The Monetary Authority of Singapore is working alongside financial institutions on runtime safeguards to verify agent authority prior to financial transactions, while IMDA collaborates with tech firms like Google and Cloudflare to explore adapting the internet’s Domain Name System for cross-platform agent verification. Singapore also intends to use its position as ASEAN chair in 2027 to advance regional standards in AI testing and governance.
Agents can act for us, but they cannot answer for us. Our task is to make sure that behind every agent, there is always someone who does.
Jasmin Lau, Senior Minister of State for Digital Development and Information
As Singapore continues to shape these digital frameworks, officials emphasize that maintaining clear accountability remains essential for the safe integration of artificial intelligence into everyday financial and professional activities.
