Meta Platforms Inc. is partnering with Walmart Inc., Stripe Inc. and the enterprise artificial intelligence firm Sierra Technologies Inc. on the creation of new standards that will define how autonomous AI agents interact with other businesses online.
The new standard was introduced by Sierra co-founder Bret Taylor, the former co-chief executive of Salesforce Inc., who said it will help businesses to distinguish purchases from AI agents that are authorized to work on behalf of their users. “We’re designing it to handle authentication, empower consumers and give companies visibility into what personal agents do through their websites, APIs or company agents,” Taylor wrote in a blog post. “It’s open for anyone to implement.”
Many AI companies, including Meta with its personal AI agent Muse, are betting big on agentic commerce, or the idea that people will simply use agents to make purchases for them online. However, some major concerns have been raised about the safety of AI agents and the issues that could arise from giving them access to personal funds and letting them loose on the internet.
“We’re defining rails that we hope personal agents and business agents can run over for the future,” David Singleton, Meta’s Superintelligence Labs vice president of engineering and consumer products, said in an interview with CNBC. “If you think about email, it works great because it is a standard that everyone can use to talk to each other, and we think this is going to be quite similar.”
Meta, Sierra and their partners call the new standard the Personal Agent Protocol, and it appears to be a response to a call from six major banks, including Bank of America and Capital One, to provide more clarity around AI agents. Last month, the banks published a paper that urged the AI industry to establish comprehensive standards, policies and consumer protections based on five key principles: transparency, safety, privacy and data, choice and interoperability.
The fears around AI agents are widespread. Some fear that their adoption could lead to a rise in financial scams, fraud and disputes, the banks claimed. They also believe that AI agents could prioritize certain kinds of products or payment methods due to incentives such as higher commissions, rather than choosing what’s best for their users.
Meta’s new agent Muse, which only launched last month but has already become one of the most downloaded applications in Apple Inc.’s App Store, has already raised privacy concerns. On Monday, 404 Media reported that Meta’s engineers had to race to fix multiple dangerous security vulnerabilities in the agent to get it ready for its launch date. But even with Meta working overtime, researchers have already raised a number of privacy and security concerns over Muse.
Muse was notably also banned from doing anything on Amazon.com Inc.’s online retail platform, after that company said the agents failed to identify themselves, leading to concerns over how they might handle user’s account data and credentials. This week, Muse hit the headlines again when Wired reported that Muse was creating profiles for every single person in its user’s lives, even if they don’t use Muse themselves.
Taylor, who appears to be at the forefront of the Personal Agent Protocol’s development, told CNBC that there “will be chaos until such a standard exists” that can ensure visibility and transparency into what personal agents are doing on retailer’s websites. Taylor, who is also chairman of OpenAI Group PBC, said he expects that the AI giant will also back the standard in the future.
Singleton said Meta hopes the standard will make commerce agents easier for consumers to adopt, trust and use. Meta has good reason to want this to happen. According to a report from Citigroup earlier this week, Muse could generate upwards of $27 billion in revenue for the company by the end of the decade as it pushes to become the “front door” to the world of online shopping.