Meta built an AI that can shop for you. The problem is that most people don’t want AI spending their money

After a long week at work, the perfect Saturday for some people involves grabbing coffee with friends and walking around, spending the hard-earned money they toiled all week to earn. Shopping can be a form of therapy—it’s something even Cher Horowitz understood in Clueless. But Silicon Valley has a different idea: Shopping is a chore best handed to artificial intelligence.

On Tuesday, Meta became the latest tech giant to join the AI shopping rush, with Muse: a personal agent that can use years of social activity to understand a consumer’s taste, search for products, navigate checkout, and prepare a purchase for final approval. It joins a rapidly expanding crowd of AI personal shoppers from Amazon, Google, Walmart, Disney, Uber, and OpenAI, all competing to control the increasingly valuable journey between wanting something and paying for it.

Muse, which will initially roll out to U.S. adults for free, can act autonomously on a user’s behalf to complete tasks, ranging from messaging friends and planning a night out to shopping for a vacation outfit.  Unlike a conventional chatbot that suggests products or compares prices, Muse can navigate checkout and pay through Stripe’s Link service after presenting the total for the consumer to approve, according to Meta.

But the industry’s eagerness to automate shopping is running ahead of consumers’ willingness to actually surrender it. Some consumers are already embracing AI for parts of the shopping process: traffic from AI-powered assistants and chatbots to retail websites jumped 693.4% during the 2025 holiday season, according to Reuters. Online holiday spending reached a record of $257.8 billion during the period, up 6.8% from the previous year. 

However, that may be the extent of how much consumers want AI in their shopping cart. Only 31% of respondents to a Vogue Business survey said they would outsource shopping to an AI agent, even if it understood their taste and purchase history. Just 24% trusted recommendations and summaries produced by AI chatbots, while 72% said they would not share card details, 46% would withhold their browsing history, and 40% would not provide location data.

More than half of the respondents had never used AI to shop for fashion or beauty, and only 2% said it consistently understood their personal style. Consumers were more receptive to using AI as an assistant than allowing it to become the buyer, with some worried that making purchases too automatic could cause them to lose control of their spending—and that’s because asking AI to find the cheapest TV is different than letting it pick and choose one for you.

Meta’s data advantage comes with privacy baggage

Shopping is personal not only because people enjoy doing it, but because what they browse, want, and ultimately buy can say a lot about who they are.

Most shoppers appear more interested in using AI to find deals or narrow their options than allowing agents to choose products and spend their money. Meta, however, has access to years of information about what people watch, follow, save, and share—giving it a window into not only what consumers might buy, but who they are.

Muse can remember information a user shared once and turn a recipe reel saved on Instagram into a grocery list, according to Meta. The memory could make Muse an unusually perceptive shopper, but it requires access to increasingly personal information. Meta says users choose which apps Muse can connect to and how much access it receives, can disconnect services at any time, and can opt out of having their interactions used to train its AI models. The company also says Muse conversations and data will not be shared with its advertising systems.

Those safeguards must overcome Meta’s long history of backlash over how it collects and uses personal information, including a record $5 billion Federal Trade Commission penalty in 2019 over allegations that Facebook violated users’ privacy. Muse is now asking consumers to grant Meta access to even more personal information in exchange for better recommendations. The more Muse knows, the more useful it could become—but trusting Meta to turn personal data into purchases is a much bigger ask.

Muse also represents a consumer-facing test of CEO Mark Zuckerberg’s costly effort to put Meta back at the front of the AI race. After its previous model, Llama 4, was widely panned, Meta spent $14.3 billion for a 49% stake in Scale AI, hired its cofounder Alexandr Wang as chief AI officer, and committed hundreds of billions of dollars to AI infrastructure. Muse runs on Muse Spark, which returned Meta to the race without consistently beating models from OpenAI, Anthropic, and Google.

Placing Muse across Meta’s social empire gives that investment a direct route into consumers’ everyday lives—and potentially the moment they open their wallets. But Meta’s ability to put another AI product in front of billions of people does not mean they want it there.

Regardless, Meta’s not the only one entering the shopping space. In May, Amazon introduced Alexa for Shopping, while Disney began testing a conversational AI personal shopping assistant inside its store app in June. Uber Eats’ Cart Assistant can turn a typed prompt or photograph of a handwritten list into a grocery cart informed by past orders. By 2030, Mastercard predicts one in 10 shoppers will use a personal AI agent to purchase products or services.

But getting there may require tech companies to recognize that shopping is not always a problem people want solved. Consumers may welcome help comparing thousands of products or finishing a tedious grocery order. Browsing without a plan, stumbling onto something unexpected, and choosing for themselves are the parts they may want to keep. The winner of the AI shopping race may be the company that understands when consumers want an agent to take over—and when they merely want someone to hold the bags.

This story was originally featured on Fortune.com