Trending
Meta has launched its long-awaited AI assistant, Muse, which can perform tasks on behalf of users across several digital services.
The company says Muse can send emails, sell a car and book travel with user permission. The launch forms part of CEO Mark Zuckerberg’s broader push to develop what he calls “personal superintelligence.”
Meta is also seeking new revenue streams as it increases spending on artificial intelligence. The company expects its investment in AI chips and related infrastructure to exceed $130 billion this year.
Muse will initially be available only in the United States. Users can access it through a dedicated app or Meta’s WhatsApp service.
Meta also plans to bring the agent to its smart glasses soon.
A basic version of Muse will be available at no cost.
Meta will also offer paid subscriptions for users who require heavier access. The plans will cost $20 and $100 per month, according to a company spokesperson.
Users can choose whether their interactions with Muse can help train Meta’s AI models.
KMEDA Helps Recover Customs-Held Mobile Phones From TCS Shipments
The company also plans to introduce an encrypted version of the service later this year.
Muse is designed to connect with apps covering email, calendars, payments, health, shopping and smart-home services.
Users decide which applications Muse can access. They can also revoke that access at any time.
Each Muse agent operates inside its own virtual machine. This allows the system to continue working in the background even when users are not actively interacting with it.
The wider access could make Muse more useful. However, it also creates greater security risks if the agent makes an unintended decision or mishandles personal information.
Meta Vice President of AI Products Vishal Shah said the company delayed the product’s release in April to improve security.
Shah told Reuters that the additional work helped Meta reach what it considered the minimum safety threshold for a public launch.
Despite Meta’s assurances, internal testing has produced mixed results, according to posts reviewed by Reuters.
One employee said Muse had become highly useful for organising travel during a three-week honeymoon in Indonesia.
Other employees reported more serious problems.
One internal test reportedly showed an agent bypassing safeguards and exposing personal iCloud photos while searching for toys in images from a child’s birthday party.
Meta Chief Technology Officer Andrew Bosworth also reported repeated login problems while testing the product.
Another employee said Muse struggled to monitor items that sell out quickly. The employee reported several failures, including the system stopping page refreshes and disabling monitoring without an obvious reason.
Meta did not respond to Reuters’ request for comment about the specific incidents described in the internal posts.
The Muse launch comes as technology companies face growing concerns about autonomous AI agents.
AI systems developed by several major laboratories, including OpenAI and Anthropic, have faced problems involving unexpected behaviour and attempts to work around restrictions.
Meta has also experienced a rise in technical and security incidents linked to its growing use of AI coding tools and autonomous agents.
According to internal figures cited by Reuters, major incidents at Meta increased by 40% from last year. Staff time spent dealing with those problems also rose by 70%.
The figures highlight the challenge facing Meta as it expands autonomous AI systems that can interact with real-world accounts, services and personal data.
Copyright © 2026 99News. All Rights Reserved.
