Singapore is exploring whether the internet’s domain-name system can help identify AI agents and verify who controls them. The practical question is whether existing infrastructure can make autonomous software more accountable across organisations, without building a registry from scratch.
Watch Desk analysis
What happened
Singapore’s Infocomm Media Development Authority is working with GoDaddy, Infoblox, Cloudflare and Google on ways to adapt DNS for identifying and verifying AI agents, according to Biometric Update, which cites The Straits Times. The collaboration has produced a demonstration: in October, the companies presented a “Chain of Trust” architecture intended to verify agent identities during transactions across independent platforms. Read the report.
A separate proposal from MP Tin Pei Ling would create a national registry linking agents to accountable organisations and authorised activities. Meanwhile, Singapore’s GovTech agency is developing GovHarness, shared infrastructure for building and operating AI agents across the public service. These are related efforts, not evidence that a national registry or a DNS-based identity system has been adopted.
Why it matters
Agents that act across services raise a simple accountability problem: when software takes an action, how can another system establish which agent it was and who is responsible for it? A shared identity layer could make verification more practical across platforms, while giving organisations a clearer way to connect agents to their operators.
That promise comes with a design question worth watching. An identity mechanism needs to be interoperable enough to work across services, but it should not quietly become a universal tracking system. The report describes exploration and a demonstration, not a deployed standard.
Our read
This is a useful move from abstract arguments about “responsible agents” towards the infrastructure that might make responsibility traceable. Building on DNS is an intriguing route; it is also only a starting point. The important test is whether the system can verify an agent and its authority without turning every interaction into a new layer of surveillance.
What to watch
- Whether IMDA publishes technical details of the Chain of Trust demonstration.
- Whether the companies propose an open standard or a production deployment.
- How agent identity would connect to operator accountability and authorised actions.
- Whether GovHarness and the DNS work remain separate or become part of a shared government approach.
Discussion spark: Should AI agents have a shared, verifiable identity across platforms, even if that makes their actions easier to track?
Sources and evidence
- Source update (13 November 2026, 19:50 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.