Vint Cerf calls for internet standards to identify AI agents

Vint Cerf is calling for formal internet standards that can identify autonomous AI agents as they move across the open web. The internet architect, who co-designed TCP/IP with Robert Kahn, says agents from different companies will need a common language and a trustworthy identity layer if they are going to browse, negotiate and act online on behalf of people.

The current web, built around human users, is colliding with software that can operate continuously and at machine speed across many services. That shift raises fresh risks around trust, authentication, discovery and abuse, especially if agents are allowed to transact or make decisions without a standard way to tell who built them, what they can do and whether they are authorized.

On June 23, 2026, the Linux Foundation announced plans to launch the Agent Name Service, or ANS, an open standard for trusted identity, verification and discovery for AI agents operating across the internet. NIST also launched an AI Agent Standards Initiative in 2026 to promote industry-led technical standards and open protocols for autonomous agents.

Google's open Agent2Agent protocol is designed to complement Anthropic’s Model Context Protocol, while Agentic Resource Discovery is an open specification for publishing, discovering and verifying AI capabilities across the web. Google has also shipped agent-building tools including Gemini API Managed Agents, ADK 2.0, Antigravity 2.0 and Genkit Agents.

He had been with Google since October 2005, serving as vice president and chief internet evangelist for roughly 20 years. The retirement was announced to attendees at the Open Frontier conference hosted by the Laude Institute, where Dave Patterson announced it to attendees.