Key facts
- Personal AI agents are facing website blocks, hindering tasks like booking flights and ordering groceries.
- Amazon has blocked Meta's Muse AI agent from its retail site.
- Walmart stated its website blocks were not intentional and that it is partnered with Muse.
- Meta, Walmart, Stripe, and other industry partners are developing an open standard for AI agent communication.
- Airlines like Delta and United are implementing measures to protect against unauthorized automated activity.
- Yelp requires agents to pay for data access to use its platform.
Personal AI agents, designed to perform tasks such as booking flights and ordering groceries on behalf of users, are encountering significant obstacles as many websites are blocking their access. This issue is frustrating consumers and highlighting a potential incompatibility between current anti-bot measures and the emerging era of agent-first web interactions.
Companies like Amazon have begun actively blocking AI agents, such as Meta's Muse, from their platforms, preventing them from browsing or making purchases. While some blocks are intentional, others stem from traditional anti-spam and anti-malicious activity measures that inadvertently prevent legitimate AI agents from operating. The lack of clarity for end-users on whether a block is intentional or accidental adds to the frustration.
Walmart, despite facing user complaints about Muse being blocked, stated that these were not intentional and that the company is partnered with Muse. A spokesperson explained that website checks designed to verify human visitors can sometimes interrupt AI agents, leading to failed transactions. This suggests that current methods for distinguishing humans from bots may need to be updated for the agent-first web.
To address these challenges, industry partners including Meta, Walmart, Stripe, and others are collaborating on an open standard. This protocol aims to define how AI agents can communicate with businesses, specifically for online commerce, to differentiate between beneficial agents acting on behalf of users and malicious bots. Meanwhile, other companies, such as airlines like Delta and United, are taking a cautious approach, implementing security measures to protect against unauthorized automated activity and emphasizing the need for security and customer experience to be considered if they were to integrate with AI agents.
Some platforms, like Yelp, have explicit policies requiring agents to pay for data access, while eBay restricts unauthorized agents and scraping. Cloudflare, a provider of website protection, has also made changes to its systems that may be disrupting AI agent traffic, though the company has not provided specific data on this issue. Meta's strategy of forming specific brand partnerships and focusing on new standards is seen as a way to ensure its agent, Muse, is welcomed by partners and to facilitate reporting of any access issues.