After 20 years at Google, the 'father of the internet' is retiring

The internet as we know it would look very different without Vint Cerf.

Decades before the web became an indispensable part of everyday life, Cerf helped develop the networking protocols that laid its foundation. Now, after a career spanning more than half a century and over 20 years at Google, the 83-year-old computer scientist, widely recognised as one of the "fathers of the internet", is preparing to retire from his role as the company's Chief Internet Evangelist next week.

The announcement was made during the Open Frontier conference hosted by the Laude Institute, where fellow computer scientist and UC Berkeley professor Dave Patterson paused a panel discussion to recognise Cerf's extraordinary career.

"Vint... has been at Google more than 20 years, and he is retiring a week from today, and so I think we ought to give him a round of applause for a relatively good career," Patterson said, prompting applause from the audience.

More from Tech

Even as he prepares to step away from Google, Cerf remains focused on what he believes could become the next defining technological challenge: ensuring artificial intelligence develops on open, interoperable standards much like the internet did.

Vint Cerf: The man who helped build the internet is now thinking about AI

Cerf's influence on computing stretches back to the 1970s, when he and fellow computer scientist Robert Kahn developed TCP/IP, the networking protocol suite that became the foundation of today's internet.

That work earned Cerf some of the highest honours in technology, including the Turing Award, often referred to as the "Nobel Prize of Computing", as well as the US Presidential Medal of Freedom and numerous honorary degrees.

Since joining Google in 2005, Cerf has served as Vice President and Chief Internet Evangelist, a role that saw him champion internet accessibility, open standards and emerging technologies.

At the Open Frontier conference, however, his attention was firmly on the future of AI.

Speaking alongside researchers including François Chollet, John Ousterhout and Matei Zaharia, Cerf argued that the next generation of AI systems will increasingly rely on autonomous software agents interacting with one another rather than simply responding to human prompts.

"An AI agent model in which several agents from different sources interact with each other will demand composability and will require interaction and standardization," Cerf said.

The observation echoes the early days of the internet, when common protocols enabled computers from different manufacturers and networks to communicate seamlessly.

English is not the best choice, AI may need a language of its own, says Cerf

Cerf believes that interoperability alone will not be enough.

He suggested that natural languages such as English may prove too ambiguous for AI agents expected to cooperate without human intervention.

"I don't think English will be the best choice," Cerf said. "It is flexible, but it also has ambiguity, and precision in inter-agent interaction will be incredibly important. An agent must be confident that the other agent understands exactly what they agreed to do together."

To illustrate the problem, Cerf compared AI communication to the children's game of telephone, where a whispered message gradually changes as it passes from one person to another.

"Remember the old game of telephone, where you wanted to whisper something into someone's ear, but by the time ten people later the message came out different?" he said. "Imagine a group of agents that speak in natural language. It's a bit scary."

The comments come as technology companies race to build AI agents capable of performing complex tasks independently, making questions around interoperability and standardisation increasingly important.

Patterson also offered a lighter reflection on Cerf's legacy, joking that beyond his technical achievements, "He has always been the best-dressed computer scientist I've ever met."

As Cerf retires from Google, the internet pioneer leaves behind more than an extraordinary career. His work helped establish the open standards that connected the world. Now, his final message to the AI industry is strikingly familiar: the future will depend not just on smarter machines, but on ensuring they can understand one another.