Meta and Partners Launch Global Standard to Regulate and Secure AI Agents
A Tech Alliance to Tackle the "Chaos" of Digital Agents
Meta has announced a joint collaboration with a group of leading global companies—most notably Walmart, Stripe, and AI startup Sierra—to develop an open technical standard designed to unify how AI agents interact with businesses and commercial services, while bolstering data security and login authentication.
"Personal Agent Protocol": Digital Identity and Greater Transparency
The new standard, named the Personal Agent Protocol, aims to verify the identity of digital agents and define their authorization scopes when performing tasks on behalf of users. Bret Taylor, co-founder of Sierra, former co-CEO of Salesforce, and chairman of OpenAI's board, stated that the lack of unified standards creates a state of "chaos." He emphasized the urgent need for enterprises to distinguish between authorized personal agents acting with user consent and unauthorized bots attempting to access systems.
Protecting Sensitive Data and Enhancing Transaction Security
David Singleton, VP of Engineering at Meta Superintelligence Labs, stressed that the primary objective of the new standard is to enhance security and transparency—especially as smart agents handle sensitive information such as payment card details. This effort comes amid a rapid rise in the deployment of AI agents for shopping, task execution, and daily digital interactions.
Automated Access Limits and E-Commerce Disputes
The launch of this protocol comes at a time of fierce competition among tech giants and growing legal challenges regarding unauthorized automated scraping and access. Major players like Amazon have recently taken blocking measures against certain Meta AI agents, while other companies face similar legal disputes over agent deployment (such as lawsuits involving Perplexity). The protocol seeks to establish a delicate balance between simplifying automated tasks and maintaining robust system security.