Every few years, a new technology captures the world's imagination and prompts the same question across every industry: "How does this affect what you do?" Today, that technology is artificial intelligence (AI), specifically large language models (LLMs) like ChatGPT and its counterparts. At ICANN, we hear this question frequently. How does AI affect the Domain Name System (DNS), the identifier ecosystem, or ICANN's mission?
AI does not change ICANN's core mission or the fundamental architecture of the DNS. But like any significant technology shift, it does have implications for the ecosystem ICANN operates in, and some of those implications deserve careful attention.
"AI" has become the dominant technology narrative, much as "blockchain" was a few years ago, and "big data" before that. The arrival of a widely discussed new technology brings curiosity, and can also bring a degree of hype that makes it difficult to distinguish the potential for meaningful change from general enthusiasm.
In the case of ICANN's core mission (helping ensure a stable, secure, and unified global Internet by coordinating the Internet's unique identifiers, including domain names, IP addresses, and protocol parameters), AI does not change the fundamentals. The identifier layer of the Internet is infrastructure. It does not generate content, make decisions, or interact with users in the ways that LLMs do. At a basic level, AI systems rely on the DNS the same way as every other Internet application does. But the scale and nature of that reliance is worth examining.
AI and Changes in How the DNS is Used
One observable shift is in DNS traffic itself. LLMs and agentic AI systems interact with the Internet differently than human users do. When a chatbot searches the web to answer a question, or an AI agent performs research, gathers data, or takes actions across multiple websites, each of those interactions generates DNS queries. These machine-driven queries can differ from human browsing patterns in volume, frequency, and distribution. While the DNS was designed to handle growth and changing usage patterns, the emergence of AI as a significant source of Internet traffic is a development worth monitoring.
A more pressing concern is the use of AI to scale malicious activity. LLMs can make it significantly easier to generate convincing phishing content, create impersonation sites, or craft targeted social engineering campaigns. Many of these activities depend on domain names. AI tools can also be used to register domains for abuse campaigns at scale, and to do so in ways that may be harder to detect using traditional methods. The potential for AI to accelerate and automate domain name abuse represents a meaningful shift in the threat landscape.
On a positive note, other AI technology can be used on the defense side to detect this malicious activity: machine learning is a type of AI where computers learn to recognize patterns and make decisions from data, rather than being explicitly programmed with rules. ICANN's own research teams use machine learning to cluster domain registrations, proactively detect malicious domains, and analyze abuse patterns at scale.
AI and ICANN's Multistakeholder Processes
There is another dimension that goes beyond technical infrastructure. ICANN's legitimacy rests in significant part on the integrity of its multistakeholder model, which depends on genuine human participation in policy development. AI tools make it possible to generate public comments, draft policy language, respond on mailing lists, and engage in community processes at a scale and speed that could conceivably affect those processes. The question of how to detect AI-generated participation, and how to weigh it, is a governance challenge that ICANN, along with many other organizations, will likely need to confront.
ICANN has a specific stake in the reliability of AI-generated information because part of its mission involves being a source of truth. Through the IANA functions, ICANN maintains authoritative registries of protocol parameters, number resources, and top-level domains. These are datasets where precision matters. They are not the kind of information that can be casually summarized without loss of fidelity. When LLMs intermediate access to authoritative data, they can act as imperfect narrators, introducing inaccuracies or losing critical precision. While this risk is not unique to ICANN, it is particularly relevant to any organization whose core function includes disseminating data that must retain its exactness.
The Bottom Line
When a significant new technology comes along, ICANN, like any other organization, needs to understand its effects. Specifically in the case of AI, the technology does not change ICANN's mission, but it does change the environment in which that mission is carried out. AI has the potential to reshape DNS traffic patterns, scale up both abuse and defense, raise new considerations about the integrity of participation, and highlight the importance of ICANN's role as a source of authoritative data. These are not reasons for alarm, but they are reasons for ICANN to follow these AI-related developments closely.
