Tuesday, 8 September 2026
Rīga TV

World and Latvian news in one place

TechnologyPublished: 8 September 2026 at 22:09

Meta unveils Muse, its AI agent that can shop, email, and plan trips

Meta has introduced its long-awaited personal AI agent Muse, which can handle online shopping, email drafting and sending, and trip planning on behalf of users. The agent is designed to work independently, but requires access to users' digital accounts and services.

Foto: Engadget

Meta has finally revealed Muse, its personal AI agent that the company says can be set up without any technical expertise. The agent is meant to perform various tasks on a person's behalf, including online shopping, composing and sending emails, and planning trips. Unlike the existing Meta AI assistant, Muse is positioned as a more capable product with dedicated apps and a website, and it is able to tackle complex jobs and work autonomously.

According to Meta, once a user shares a goal, Muse helps create a personalized plan and then coordinates time and resources to push the work forward on its own. What Muse can actually do depends on how much of a person's digital life they allow it to access. It can connect to Meta's own apps and third-party services such as Google Workspace, Ticketmaster, OpenTable, Spotify and Apple Health. For instance, a user might ask Muse to go through recipe videos saved on Instagram or Facebook and organise them into a more readable format, then use its shopping and browsing tools to build a grocery list, compare prices and place an order.

Meta's VP of AI Products, Vishal Shah, told Engadget that because Muse is a capable AI product that can also build its own software, it is not really limited in what it can do. The catch is that users need to feel comfortable giving it more access and explaining what matters to them. To prevent mishaps, Meta says it has built in safeguards. Muse will ask for approval before sending an email or completing a purchase, while 'low-risk' tasks may be handled without sign-off. The agent runs on a dedicated virtual machine designed to keep each person's agent and data separate from others'. Meta acknowledges that Muse can still make mistakes, but expects them to be less frequent and less damaging thanks to built-in safety systems.

Zuckerberg has long framed AI agents as a step toward bringing 'superintelligence' to the general public, and Meta is already integrating Muse into WhatsApp and its smart glasses. Yet privacy concerns remain. The company recently faced criticism over its AI glasses and a feature that generated AI images from public Instagram posts, which was later rolled back. Meta is allowing users to opt out of having their interactions with Muse used for further AI training, and says it will not share details from conversations with its advertising systems, though shopping interactions could indirectly influence ads. It is also unclear when Muse might require a paid subscription; Meta says most of what people need will stay free, but doing more would cost.

Comments

0/1500

Comments are automatically moderated. No hate, threats, personal data or spam.

Loading comments…

More in this category