Define Gravity - Managed Growth Infrastructure
Get a quote

Governance at Scale vs Truth at Scale: What ICANN's AI Notes Imply for Operators and Brand DNS

ICANN recently published two posts that are worth reading together. One is about artificial intelligence and what it means for DNS, abuse, multistakeholder work, and how people use authoritative ICANN data. The other is about moving toward the opening of the 2026 new gTLD round, including timing and application-system readiness.

First thread: AI and the work of ICANN

Matt Larson's post, Artificial Intelligence and the Work of ICANN (24 March 2026), states that AI does not rewrite ICANN's core mission or the basic architecture of the DNS. Identifiers stay infrastructure. What changes is the environment: more machine-driven DNS use, scaled abuse that still depends on domain names (including phishing and registration behavior), and defensive uses of techniques such as machine learning to detect patterns. It also raises governance questions when policy participation can be generated at volume, and it warns that summaries of authoritative ICANN data are a poor substitute for exact primary material. For operators and advisors, that is one coherent read: noisier traffic, harder trust, same requirement to read the source.

Second thread: the 2026 new gTLD round

Theresa Swinehart's post, Advancing Toward the Opening of the 2026 Round (25 March 2026), covers readiness for the next application round: application-system work, security testing materials, Community Priority Evaluation for contention cases, and resources for applicants. It calls out 30 April 2026 as the opening date for applications.

The 2026 round is governed by the Applicant Guidebook. Application requirements, evaluation, and contention are spelled out there. The current English guidebook is available as a PDF.

Why both belong in one notebook

They are not the same story. The first describes a noisier environment around stable identifiers. The second opens a specific window for organizations that may apply to run a new top-level domain.

For brands that are ready to treat a namespace as long-term infrastructure, the second thread is the unusual part: a chance to tighten how official is defined in your own slice of the DNS, with registry-level control and real operational work behind it. The first thread is the reminder that influence and fraud both got cheaper to scale.

A lot of the AI conversation is noise: fear, hype, and vendor fog. If your organization operates an approved gTLD, you are the registry for that top-level domain: you set policy for names under your string within ICANN rules and your contracts. That is not the same as controlling the whole internet, the DNS root, or beating AI. Phishing and impersonation still happen on other domains and channels. The honest upside for many brands is clearer communications and a contract-backed home for the journeys and properties you place on that TLD, with room to be creative inside rules you help run. It does not automatically make every risk disappear; it gives you a defined place where your issuance and your operations can stay explicit.

Sources