Add as a preferred source on Google

GoDaddy Legal Chief Says AI Agents Need a Shared Identity Standard

Jared Sine says Agent Name Service could help businesses check who operates an agent, what it is authorized to do and whether its software or ownership has changed.

Two unmarked metal identity tags rest beside a network server plaque / TokenPost.ai
Two unmarked metal identity tags rest beside a network server plaque / TokenPost.ai

GoDaddy Chief Strategy and Legal Officer Jared Sine argues that AI agents need a shared way to verify identities and permissions as they begin moving money, sharing data and completing transactions online.

GoDaddy contributed to Agent Name Service (ANS), which is now a Linux Foundation project. Sine says the standard would give agents a public way to identify one another and check who operates them and what they are authorized to do.

The system could also help businesses verify whether an agent’s software matches its registered version and whether its ownership or software has changed. Sine says the Domain Name System (DNS), which connects website names with internet addresses, can adapt to support agent identities.

Sine argues that identity checks matter beyond preventing impostors. In one scenario he describes, a supplier could buy a company’s procurement agent while keeping its software and domain name unchanged, then influence the agent to recommend the supplier’s products. The business using the agent might not see that its incentives had shifted.

A shared verification layer, Sine says, could help businesses detect changes to an agent’s operator, software or ownership. He presents that as a way to check whether the agent is still acting as expected.

Sine also warns that a patchwork of proprietary protocols could keep agents inside private commercial networks. He argues that a shared standard would support broader interoperability, similar to the open infrastructure that helped the internet develop.

ANS is intended to provide that common identity layer for AI agents.

Loading…