← 返回新闻首页

'Who answers for AI agents?' Singapore working on safeguards as adoption grows, says Jasmin Lau

Singapore is working with industry, regulators and international partners to develop safeguards and standards that can support the wider adoption of artificial intelligence (AI) agents, said Senior Minister of State for Digital Development and Information Jasmin Lau in Parliament on Wednesday (Nov 4).Marine Parade-Braddell Heights MP Tin Pei Ling had raised an Adjournment Motion on a national digital identity infrastructure...

AsiaOne Singapore2026-11-04 21:00查看原文 ↗

'Who answers for AI agents?' Singapore working on safeguards as adoption grows, says Jasmin Lau

Marine Parade-Braddell Heights MP Tin Pei Ling had raised an Adjournment Motion on a national digital identity infrastructure for AI agents.

She proposed a registry that would allow different parties to verify an agent's identity and mandate.

Responding to Tin, Lau said trust is crucial as AI agents become increasingly capable of carrying out tasks on behalf of individuals and organisations.

Unlike conventional AI applications that mainly generate information or recommendations, AI agents can interact with different systems and complete increasingly complex tasks with varying levels of autonomy.

Lau said the Government supports the objectives behind the proposal and wants to ensure AI agents can operate securely across different platforms and organisations without creating separate "walled gardens", as referenced by Tin.

According to Lau, confidence in AI agents depends on three key questions being answered: who the agent is, what it is authorised to do and who is accountable for its actions.

These safeguards are especially important as AI agents begin handling more consequential tasks, such as financial transactions and access to sensitive information, she added.

She pointed to efforts already underway by merchants in the private e-commerce and payments sector, including Visa and Google.

In financial services, the Monetary Authority of Singapore is developing the Safeguards for Agentic Finance at Runtime with financial institutions and AI firms.

The system is designed to verify an agent's identity and authority before transactions are executed, while keeping auditable records of its actions.

Separately, the Infocomm Media Development Authority (IMDA) is working with GoDaddy, Infoblox, Cloudflare and Google to explore whether the internet's Domain Name System (DNS) can be adapted to identify and verify AI agents across different platforms.

Within the public sector, GovTech is also developing an internal registry to help agencies track the AI agents they deploy and the officers responsible for them.

Lau noted that countries including China, the US and Estonia are also exploring ways to establish trusted identities and governance frameworks for AI agents.

As ASEAN chair in 2027, Singapore intends to strengthen regional capabilities in AI governance, testing and assurance, including the adoption of international standards and common testing tools for agentic AI.

The Government's approach, Lau said, is to work with industry, build on emerging standards and assess whether additional common infrastructure, such as a national registry, may eventually be needed.

"We share Ms Tin's ambition for Singapore to be a place where people and businesses can use AI agents with confidence," she said.

Lau concluded saying: "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."

laili.abdeen@asiaone.com

新加坡数码发展及新闻部高级政务部长刘嘉敏周三（11月4日）在国会表示，新加坡正与业界、监管机构和国际合作伙伴共同制定保障措施和标准，以支持人工智能（AI）代理的更广泛应用。

马林百列-布莱德尔高地国会议员陈佩玲提出了一项关于为人工智能代理建立国家数字身份基础设施的休会动议。

她提议建立一个登记册，允许不同方核实代理人的身份和授权。

刘先生回应丁先生的观点时表示，随着人工智能代理越来越能够代表个人和组织执行任务，信任至关重要。

与主要生成信息或建议的传统人工智能应用不同，人工智能代理可以与不同的系统进行交互，并以不同程度的自主性完成越来越复杂的任务。

刘先生表示，政府支持该提案背后的目标，并希望确保人工智能代理能够在不同的平台和组织中安全运行，而不会像田先生提到的那样，创建单独的“围墙花园”。

Lau认为，对人工智能代理的信心取决于三个关键问题的答案：代理是谁，它被授权做什么，以及谁对其行为负责。

她补充说，随着人工智能代理开始处理更重要的任务，例如金融交易和访问敏感信息，这些保障措施尤为重要。

她指出，包括 Visa 和 Google 在内的私营电子商务和支付领域的商家已经在努力解决这个问题。

在金融服务领域，新加坡金融管理局正与金融机构和人工智能公司共同制定运行时代理金融保障措施。

该系统旨在交易执行前验证代理人的身份和权限，同时保留其操作的可审计记录。