Meta Scales Back AI Calling Feature After Human Operator Concerns Meta rolled back an AI calling feature that used human contractors after internal concerns over privacy and a racist remark, Reuters reports. The Muse feature is under review. Meta has rolled back a feature of its AI calling assistant, known as Muse, after internal concerns arose about the use of human contractors to make calls on behalf of users, according to a Reuters report. The feature, tested by employees, allowed a human concierge to step in when the AI voice failed. The rollback was confirmed by a Meta spokesperson and follows reports of privacy worries and a racist remark by a human operator. Meta AI Calling Feature Under ReviewReuters reported last week that Meta had been testing AI-powered calls that could handle tasks such as booking haircuts, checking store inventory, or getting quotes from contractors. The report said some Muse tasks involved a human concierge feature, where a human operator would take over when the AI voice was not up to the task.Internal conversations at Meta, apparently seen by Reuters, shed light on why the human feature was deemed necessary. One user said that when he tried to get the Muse bot to call his insurance company, the company repeatedly hung up when the bot disclosed it was not human.When the human calling feature was added for Meta employee testers, they were allowed to opt out. But internal conversations also questioned the wisdom of the feature, with some employees worried it could feed sensitive user information to a human operator.Reuters described the feature as able to handle errands like booking haircuts, checking whether a store has an item in stock, or getting quotes from different contractors. While it may not be invasive if a human contractor calls a barber, it could be invasive if they called a urologist, for example.Privacy and Racism Concerns Prompt RollbackOne employee who tested the human calling feature reviewed a transcript of a call and claimed that the human operator involved had made a racist reference, according to Reuters. A Meta Superintelligence Labs vice president reportedly apologized to the employee who complained and explained that the company was no longer working with that individual.The vice president said, per Reuters, that the company had rolled back the feature, noting that the decision to test calls from human contractors the way it had was a miss. Meta's tests, the VP claimed, show that human involvement could drive the phone call success rate up into the high 90% range instead of the lower percentage of AI calling.Gizmodo sought clarity from Meta about the lower percentage of AI calling quote. It is not completely clear from Reuters' report whether all Muse-powered calls are rolled back, or just the ones involving human contractors, which seems more likely.If calls are still happening without human contractors, it is not clear from Reuters what, if anything, is being done to compensate for AI calls being rejected. Meta spokesperson Dave Arnold provided a statement to Gizmodo that was originally given to Reuters.Arnold said that internal testing, also known as dogfooding, is core to the product development process. While the response from employees has been overwhelmingly positive, the entire point is to get feedback so we can implement safety and privacy protections and improve features before we release them publicly.He added that Meta is working with merchants to continue improving this potential calling feature and will only roll it out when it is ready and with the proper disclosures.Background on Meta AI and MuseThe rollback comes amid broader developments in AI assistants. The wearable best suited to carry a conversational AI assistant everywhere has been sitting in your ears for a decade, highlighting the growing integration of AI into daily life.Meta is expected to announce new smart glasses, upgrades to its AI models, and more at its annual developer conference, Meta Connect 2026. Live updates from the event are being followed by the Gizmodo crew.Separately, a flaw could have allowed attackers to hijack Muse and take advantage of its access to a user's apps and devices, raising additional security questions about the AI agent. The combination of privacy, security, and ethical concerns has led Meta to reassess its approach to AI calling features.Meta's decision to roll back the human concierge feature reflects the challenges of deploying AI agents that interact with the real world. The company continues to test and refine its AI offerings, aiming to balance innovation with user safety.