S’pore exploring adapting web domain name system to track AI agents, need for national registry - The Straits Times

The proposal comes amid increasing AI agent autonomy in daily work and e-commerce, requiring new ways of governance.

SINGAPORE - Authorities here are exploring whether the internet’s existing website verification systems can be repurposed to track autonomous artificial intelligence agents, as part of a broader assessment of whether a dedicated national registry is needed.

Specifically, the Infocomm Media Development Authority (IMDA) is working with technology firms to explore adapting the domain name system (DNS), often called the internet’s address book, to identify and verify AI agents across different platforms and organisations.

“Our approach is therefore to work with industry to find out what works in practice, build on existing infrastructure and emerging international standards, and step in with common standards or infrastructure where there are gaps that industry cannot fill on its own,” said Jasmin Lau, Senior Minister of State for Digital Development and Information on Nov 4.

“This will inform our assessment of whether additional common infrastructure, including a national registry, is needed,” she told Parliament.

Lau was responding to an adjournment motion tabled by Tin Pei Ling (Marine Parade-Braddell Heights GRC) on Nov 4 in Parliament, proposing that the Government consider building a national digital identity infrastructure for AI agents.

Tin suggested this take the form of an open and trusted registry against which any entity can check an agent’s identity and what is it authorised to do. It will be akin to Singpass but for AI agents.

The proposal comes amid increasing AI agent autonomy in daily work and e-commerce, requiring new ways of governance, said Tin.

DNS, the internet’s underlying directory, translates human-friendly website names into machine-readable internet protocol addresses.

The tech firms IMDA is working with are web hosting provider GoDaddy, cybersecurity firm Infoblox, web security firm Cloudflare and search giant Google.

This effort was one of several examples Lau, who is also Acting Manpower Minister, cited on how trust in AI agents is likely to be built.

“It will come from many efforts, each addressing a different part of the problem, from recognising agents at checkouts to checking their authority before a transaction, to identifying them across platforms,” she said.

Other initiatives include an internal registry the Government Technology Agency of Singapore is developing to help public agencies keep track of the AI agents they deploy and the officers responsible for them.

In the private sector, digital payments company Visa’s Trusted Agent Protocol seeks to help merchants differentiate legitimate AI agents from malicious bots, while Google’s Agent Payments Protocol seeks to enable secure payments initiated by agents acting on behalf of users.

The call for better governance also follows many recent high-profile incidents of AI taking actions not intended by their users.

ChatGPT maker OpenAI disclosed in July that its AI agents – bots that can carry out actions with minimal human supervision – had escaped their testing environment. The agents then breached AI software repository Hugging Face to complete a task they were given, an attack that was not intended by OpenAI.

Soon after, Anthropic, Meta and Google said their AI had also gone rogue and had hacked other organisations.

Following these incidents, leaders from the European Union and 20 countries, including Singapore, jointly called for control and international oversight of frontier AI models, a move that was welcomed by UN Secretary-General Antonio Guterres.

Since the initial declaration on Sept 21, leaders from 11 other countries have endorsed the call, including Cyprus, Switzerland, France and Sierra Leone.

Lau said that trust in AI agents, similar to how trust in individuals and organisations involved in digital transactions is established through Singpass and Corppass, is possible.

But this requires the agents’ identity, authority, and accountability to be verified across different systems.

“Agents can act for us, but they cannot answer for us,” said Lau. “Our task is to make sure that behind every agent, there is always someone who does.”