01
Safety
Safe by design.
Help agents use networks responsibly, with clear boundaries around what they can access and how they act.
Agent Networking Task Force
AI agents are already using the internet.
Let's ensure they do it safely.
The ANTF is working toward a world where AI agents use the internet safely, securely, and observably.
Explore the initiativeA shared challenge
Agents are becoming network participants.
Let’s build the foundations thoughtfully.
01The initiative
Connecting expertise.
Building common ground.
The Agent Networking Task Force is a proposed collective of engineers, researchers, and other AI professionals with a shared goal: helping AI agents use networks like the internet in a safe, secure, and observable manner.
As agents interact with more systems, the questions extend beyond any single model or application. ANTF aims to bring people together around the networking layer—and the foundations we can build together.
02Our focus
Three principles.
One connected future.
A starting point for thinking about
how agents participate in networks.
02
Security
Secure at every step.
Explore how identity, authorization, and trust can protect the interactions between agents and networked systems.
03
Observability
Observable in practice.
Make agent activity understandable, so people can trace interactions, investigate failures, and account for outcomes.
03Get involved
A shared effort
starts with people.
Building the foundations for connected agents takes different perspectives. ANTF is intended for engineers, researchers, and AI professionals interested in this shared challenge.
Participation details forthcoming
The initiative is at the proposal stage. More information about taking part will be shared here as it develops.