Meta unleashes Muse AI that can email, shop, book trips, and work for you

• Courtesy of Meta

Meta CEO Mark Zuckerberg has launched Muse, a new personal artificial intelligence agent designed to go beyond answering questions and actually perform tasks for users, from sending emails and booking trips to shopping online and making payments.

Zuckerberg said the bigger ambition is to “give every person in the world a very capable personal agent that can understand their goals and can just work on their behalf 24/7.”

Meta formally unveiled Muse on September 8, describing it as a personal AI agent that does not simply answer questions but can take over tasks and projects and turn longer-term goals into action plans.

Unlike conventional AI chatbots that primarily respond to prompts, Muse can work across connected apps and continue working even when the user is no longer actively using it. The product is a key part of Zuckerberg’s broader vision of bringing what he calls “personal superintelligence” to billions of people.

Users can ask Muse to send emails, book travel, shop online, organize schedules and work on more complex goals. It can also remember information shared by a user and use that context later, such as turning a saved Instagram recipe into a grocery list or remembering dietary restrictions while organizing a dinner.

Muse can connect with apps involving email, calendars, payments, health, shopping and smart home services. Users decide which services it can access and can revoke access at any time.

The AI runs inside what Meta calls Muse Secure VM, a dedicated virtual computer in the cloud equipped with its own browser. This allows Muse to continue carrying out tasks in the background on behalf of its user.

Meta also built a separate Sentinel agent that checks Muse’s planned actions before they reach the internet and can require the user’s permission for certain actions.

The powerful access required by an AI agent, however, brings bigger privacy and security risks.

Reuters reported that Meta initially delayed Muse’s release from April to strengthen its security. Internal testing also uncovered reliability and security problems, including instances where the agent behaved unexpectedly while accessing sensitive information. (Reuters)

Meta Vice President of AI Products Vishal Shah said it is impossible to guarantee that an AI agent will never make a mistake, but the company designed Muse’s architecture to make it as safe, secure and private as possible.

Muse is powered by Muse Spark, which Meta describes as its most capable model yet for real-world agentic work.

The distinction could mark an important shift in consumer AI. Chatbots wait for people to ask questions. AI agents such as Muse are being designed to remember, decide and act on behalf of their users.

Muse is initially available only in the United States through a dedicated Muse app and WhatsApp. Meta also plans to bring the agent to its AI glasses. A basic version is free, while heavier users can subscribe to paid tiers priced at $20 and $100 per month.

Meta has not announced when Muse will become available in the Philippines.