SINGAPORE: Consumers using personal artificial intelligence (AI) agents have been warned to be cautious about granting them access to online accounts, as scammers could manipulate the technology into transferring money or revealing sensitive information.

Unlike conventional AI chatbots, personal AI agents can carry out tasks on behalf of users, including managing emails, making bookings and accessing websites and online accounts.

However, these capabilities also expose users to privacy and security risks, according to an advisory issued on Friday (Nov 6) by the Cyber Security Agency (CSA), GovTech and the Infocomm Media Development Authority (IMDA).

Examples of AI agents include Dots, Hermes, Muse, Grok Bot and OpenClaw.

“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,” the authorities said.

Some agents can continue operating in the background even when users are not actively using them. They could also be tricked by malicious actors, websites or messages into carrying out harmful actions.

“AI agents can be very useful, but you cannot treat them like human assistants and give them the same broad access to your information and implicit trust,” the authorities warned.

Instead, users should limit the permissions they grant, retain control over important actions and understand how to revoke access.

The authorities also cautioned that users remain responsible for how they use AI agents, including actions taken on their behalf, and may be held liable depending on the circumstances and applicable laws.

LIMIT ACCESS TO SENSITIVE ACCOUNTS

Personal AI agents may request access to users’ emails, calendars, contacts, messages, files and other online accounts.

Once connected, they may be able to retrieve, process and share information from these services, including personal information belonging to others, the authorities warned.

Users should therefore connect only the apps and accounts necessary for an agent to perform its tasks, and restrict the information it can access where possible.

They should also avoid granting agents access to sensitive accounts, such as those used for banking, government or healthcare services, and refrain from sharing credit card details.

The authorities advised users not to disclose usernames, passwords, passcodes, one-time passwords (OTPs) or other authentication credentials to AI agents.

For services that use Singpass, consumers should opt for passkey authentication by default, complete the authentication process themselves and avoid approving unexpected requests.

They should also choose authentication methods that require human action when using other digital services.

Source link