Meta is testing human contractors for Muse's phone calls, with employees raising privacy concerns over sensitive information potentially reaching external agents. According to Reuters, Meta enabled the 'human concierge' feature for about half of its employees last week, allowing human contractors to handle some phone calls placed through the digital agent. Under the test, Muse can hand requests to a trained human agent, who places the call and handles the interaction on the user's behalf. The test comes shortly after Meta publicly launched a phone-calling feature for Muse, which allows the AI agent to call US businesses to handle tasks such as booking appointments, checking product availability and getting quotes.
The introduction of human agents has raised privacy concerns among some Meta employees, with internal posts showing employees warned that sensitive information could potentially be shared unintentionally with contractors working in call centres. Meta confirmed that it is testing the feature internally and said employee feedback has been 'overwhelmingly positive'. The company stated that the trial is intended to gather feedback and improve safety and privacy protections before any broader public rollout. As reported by Reuters, Meta spokesperson Daniel Roberts said: 'We're working with merchants to continue improving this potential calling feature, and will only roll it out when it's ready and with the proper disclosures.'
Amazon.com Inc. has blocked Meta Platforms Inc.'s new artificial intelligence agent from its retail site after the social media company declined a request to remove the bot. According to reports from GeekWire, Meta's Muse, which was introduced earlier this month as a task-executing personal agent, is designed to help people carry out common online tasks such as shopping and booking appointments. Amazon prohibits other companies from deploying automated tools to shop its site and started blocking Muse on Sunday night, with shoppers using Muse seeing a series of pop-ups saying its use violates Amazon's terms of use. The popup explicitly states: 'Continued access by an unauthorized AI agent violates Amazon's Conditions of Use, to which our customers have agreed.'
Amazon spokesperson Lara Hendrickson stated that the company believes third-party applications offering to make purchases on behalf of customers from other businesses should operate openly and respect service provider decisions about participation. As reported by GeekWire, Hendrickson emphasized that an opt-in requirement is standard practice for food delivery apps and online travel agencies, with agentic third-party applications such as Muse having the same obligations. Amazon has requested that Meta remove Amazon from the Muse experience, citing three specific concerns: Meta never disclosed that Muse would browse Amazon's store, the agent doesn't identify itself when it does, and it appears to capture and store customer login credentials. The company says it tried and failed to get Meta to voluntarily leave Amazon.com out of the Muse experience before taking action, with Amazon stating it is in direct conversation with Meta about the issue as of Sunday night.
The dispute centers on Amazon's claims that Muse appears to capture and store customer login credentials, while Meta counters that Muse 'has no visibility into people's passwords or payment methods' and that credentials a user shares 'go into secure storage, so Muse can use them without seeing them.' According to GeekWire, Meta launched Muse as a task-executing personal agent that plugs into email, calendar, payments, dining and shopping across iOS, Android, WhatsApp and muse.ai. The company maintains that Muse operates within secure storage protocols and doesn't have direct access to user credentials or payment information. In its Muse announcement, Meta stated that 'Muse has no visibility into people's passwords or payment methods. Any credentials a person shares go into secure storage, so Muse can use them without seeing them, including passwords a person types into the browser themselves.'
Meta launched Muse on September 8 as a general-purpose assistant that can perform multistep tasks across email, calendars, dining, payments, and shopping, reaching the top of Apple's US free-app chart within a week. The business stakes are particularly high for Amazon, which generated more than ₹5,800 crore ($68 billion) in ad revenue last year, a business that depends on people browsing its pages and seeing sponsored products. The block highlights the broader challenge of regulating AI agents across platforms, as Amazon's stance carries commercial risks - blocking ChatGPT bots could mean missing purchases from OpenAI's roughly 1 billion monthly users. Meta began testing Muse's phone-calling capability with employees in August before gradually expanding it to users following the app's debut, with the company enabling the human-concierge feature for about half of its employees last week, offering an opt-out option. The development highlights one of the challenges facing AI agents as they move beyond answering questions and begin handling real-world tasks on behalf of users.
