Meta Platforms, Sierra Technologies, Walmart, and Stripe are collaborating on the Personal Agent Protocol. This new standard aims to regulate how autonomous AI agents interact with online businesses, focusing on authentication, transparency, and consumer protection.
The initiative responds to concerns from six major banks, including Bank of America and Capital One, regarding potential fraud, privacy risks, and biased purchasing decisions by AI agents. The banks previously outlined five principles for AI safety, including interoperability and data privacy.
Bret Taylor, co-founder of Sierra and Chairman of OpenAI, stated that the protocol is open for implementation. He noted that the industry currently faces significant challenges regarding visibility into agent activities on retail websites, which the new standard intends to resolve.
Meta is integrating this standard with its AI agent, Muse. Despite its popularity in the Apple App Store, Muse has faced scrutiny over security vulnerabilities, data privacy, and its practice of creating profiles for non-users. Amazon has already restricted Muse from its platform due to identification issues.
The adoption of such standards is critical for Meta's long-term strategy. Citigroup estimates that Muse could generate over $27 billion in revenue for Meta by the end of the decade as the company aims to position the agent as a primary gateway for online shopping.
