Introduction
Autonomous AI agents are poised to handle everything from flight comparisons to routine household shopping. But as these systems begin acting on our behalf, a fundamental question emerges: how will an agent know which businesses, services, or other agents are trustworthy? The answer could determine whether the next era of the internet stays open and connected or splinters into competing, closed platforms.
What Happened
Jared Sine, chief strategy and legal officer at GoDaddy, recently argued that AI agents face the same identity challenge that once plagued the early web. Before the Domain Name System, users had to remember numerical IP addresses to reach a server. DNS replaced those numbers with human-readable names like Amazon.com and Wikipedia.org, creating stable online identities that users could trust.
GoDaddy helped develop an open standard called Agent Name Service, or ANS, now hosted by the Linux Foundation. The concept applies the DNS model to AI agents. Instead of relying on a numerical address or an unverified label, an agent could use ANS to confirm it is dealing with a legitimate airline booking system, a real merchant, or an authorized payment processor before moving money or sharing data.
Sine warns that without this layer, risks go beyond simple fraud. A business might deploy a procurement agent only to have a supplier acquire that agent, leave the software and domain unchanged, and quietly steer recommendations toward its own products. To the customer, nothing looks different, but the agent is no longer acting solely in the user's interest.
Why This Matters
ANS would let every business verify who operates an AI agent, confirm that the software matches its registered version, and detect changes in ownership or code. That verification matters in practical scenarios: a travel agent must know it is reaching an airline's genuine reservation system, not a convincing imitation. A shopping agent needs to authenticate the merchant, payment processor, and delivery provider before placing an order.
The stakes extend to the structure of the internet itself. Some companies are already building agent-to-agent communication tools for payments and logistics, but those tools often live inside commercial ecosystems. If machine identity becomes a patchwork of proprietary protocols, the agent economy could grow inside private silos where one company's agents work best with its preferred partners.
The broader internet succeeded because of shared infrastructure no single company controlled. Applying that same principle to AI agents could keep the agent economy open, interoperable, and resistant to fragmentation.
Key Takeaways
- AI agents need a trusted way to identify one another before handling transactions or sensitive data.
- Agent Name Service (ANS) adapts the Domain Name System model for autonomous software agents.
- GoDaddy helped create ANS, which is now an open project under the Linux Foundation.
- ANS can verify who operates an agent, whether its software is unchanged, and if ownership has shifted.
- Without a common identity layer, the agent economy risks splitting into closed, proprietary silos.
- Open standards like ANS aim to preserve the interoperability that defined the early internet.
Conclusion
The age of AI assistants is arriving quickly, but convenience will not matter if users cannot trust the agents acting for them. By borrowing from the lessons of DNS, Agent Name Service offers a path toward a verifiable identity layer for machines. Whether that path becomes a global standard or remains fragmented among competing platforms will shape how safely and freely the agent economy grows.

Discussion
Join the conversation
Thoughtful reactions, questions, and follow-up ideas help shape the next story.