Meta has launched Muse, a personal AI agent that can do more than answer questions. It can send emails, book trips and handle tasks across apps. But its launch comes with a big question: can users trust an AI with so much access?
Muse by Meta wants to do the work
As reported by The Times of India, Meta launched Muse in the United States on September 8. The AI can connect with email, payments, calendars, health tools, shopping and smart home systems. Users choose which apps it can access and can remove it later. The idea is simple. A user gives Muse a goal, and the agent works through the steps. It can send emails, sell a car, book travel and complete other tasks. Unlike a normal chatbot, it can work in the background through its own cloud virtual machine.
Muse is also known internally as Hatch. It is part of Mark Zuckerberg’s plan for โpersonal superintelligenceโ. The tool is available through a separate app and WhatsApp in the US. Meta also plans to bring it to its smart glasses. There are free and paid options. Meta offers a free level, while paid plans cost $20 or $100 a month, based on use. This could give the company a new way to make money beyond advertising.
Safety is the big concern
The launch is drawing attention because Muse can work with private data. It is not only reading information. It can also take action for the user. That makes mistakes more serious. Meta delayed the planned April release to improve security. The company says the work helped Muse reach the safety level needed for launch. A separate system also watches planned actions and may ask for approval before some actions are taken.
Meta says each Muse agent runs in an isolated virtual machine. Users do not have to give it access to every service. This is meant to limit what the agent can see and do. Still, internal testing has raised hard questions. Meta employees found cases where Muse exposed private iCloud photos after getting around safety limits. Other tests found the agent could stop working, ignore errors or log users out repeatedly. The company did not comment on those specific incidents. These problems have added to wider concerns about privacy, safety and reliability.
Meta makes a bigger AI bet
Muse comes at a time when Meta is facing greater public scrutiny over privacy and safety. The company is also exploring ways to earn money from the service, including commerce. The launch also follows an earlier issue around Meta’s AI image tools and Instagram photos. Meta has promised stronger safeguards and says users will control access.
The service is only available in the US for now. An encrypted version is also planned, while support for AI glasses is expected later. Muse could make everyday online tasks much easier. Yet the harder test may be trust. People may like an AI that can do things for them. They may think twice before letting it touch the most private parts of their digital lives.
Author’s Opinion
Muse shows how quickly AI is moving from answering questions to acting for people. That is exciting, but trust cannot be treated as a small detail. Meta has built strong safety checks, yet real-world use will show whether those safeguards are enough.
READ MORE: Microsoft shifts Copilot toward AI that gets work done.
ALSO READ: Tesla cybercab launch triggers NHTSA safety probe.
READ NEXT: Apple hit with ยฃ2bn UK legal fight over app tracking.
