Identity Digital's Innovation Labs division has spent 2026 building a case that the fix for AI agent accountability doesn't require new infrastructure — it can piggyback on DNS, the system that already resolves every domain on the internet. In April the division launched DNSid, described as a durable identifier that binds every AI agent to verifiable ownership. In June it formalized the idea in an IETF Internet-Draft, now published as draft-ihsanullah-dnsid-00, which specifies DNSid as a minimal identity primitive that assigns each agent a fully qualified domain name under a domain controlled by its accountable entity, and publishes a structured set of pointers in DNS TXT records to the agent's cryptographic keys, lifecycle log, and operational status. On July 15, the project picked up a notable name: Vint Cerf, who co-designed TCP/IP, joined Innovation Labs' Advisory Council to work on the framework. His stated rationale gets at the actual problem: AI agents are driving the next major architectural evolution, and questions of identity, accountability, and interoperability require the same thoughtful architectural approach that enabled the Internet to scale. The design goal is narrow on purpose. As Innovation Labs' CTO put it when the draft was submitted, the framework builds on the Internet's existing coordination layer rather than introducing a new closed or proprietary system, anchoring verifiable ownership records in DNS because it's a namespace that already operates at world scale. Forbes' coverage captured the intended function well: DNSid is meant to work as a license plate, establishing who is responsible for an agent's actions, not as a trust or reputation score. This matters because most current agent-identity work — Microsoft's Entra Agent ID, Okta's Cross App Access, the NIST/NCCoE identity initiative — solves authentication and authorization inside a single vendor's walls or a federated enterprise perimeter. DNSid is aimed at the harder, unsolved layer: agent-to-agent trust across the open internet, where there's currently no shared way to prove who owns an agent that shows up at your door. Binding that proof to DNS TXT records means any system that can already do a DNS lookup — which is to say, every system — can verify an agent's accountable entity without adopting a new protocol stack. The draft is still in early IETF review (it expires November 1, 2026, per standard Internet-Draft rules), and Cerf himself is candid that the harder questions — what authority an agent actually holds, and what commitment an organization makes when it registers one — remain open. But the approach reframes agent identity as fundamentally a DNS-layer discovery problem, not just a credentials problem, which is worth watching closely.