Paul Mockapetris, the inventor of the Domain Name System, said that artificial intelligence could become dangerous without consistent human oversight.
As the architect of a fundamental pillar of the internet, Mockapetris' perspective carries weight for developers and policymakers attempting to regulate the rapid expansion of generative AI. His warnings suggest that the current trajectory of AI development may lack the necessary safeguards to ensure long-term safety.
Speaking at the America250 × Forbes America Innovates stage, Mockapetris addressed the tech community regarding the inherent risks of the technology. He said, "AI is a powerful tool, but without human oversight it can become dangerous."
Mockapetris drew a parallel between his early work on internet infrastructure and the current state of AI. He noted that the Domain Name System was first introduced in 1983 [1] to ensure the internet could scale effectively. He said, "We built the DNS to make the internet scalable; now we need the same foresight for AI."
During the event, Mockapetris also alluded to a lack of public understanding regarding the underlying mechanisms of the technology. He said, "There’s a deep secret behind AI that most people don’t understand."
These comments were further detailed in a June 2, 2025, episode of the DomainNameWire Podcast, where Mockapetris said the importance of maintaining human control over automated systems. He said that the goal is to alert the public and the technical community to potential dangers before they manifest as systemic failures.
“AI is a powerful tool, but without human oversight it can become dangerous.”
The warning from Mockapetris highlights a tension between the speed of AI deployment and the creation of governance frameworks. By comparing AI to the 1983 launch of DNS, he suggests that the industry is currently in a formative stage where architectural decisions will determine whether the technology remains a tool or becomes an uncontrollable risk.
