Meta Launches Muse AI Agent to Send Emails, Make Payments and Handle Tasks
Synopsis
Meta has launched Muse, a personal AI agent capable of sending emails, booking travel and making purchases on users’ behalf, marking a major shift toward action-oriented AI while raising fresh questions about privacy, security and reliability.
Meta has launched Muse, a new personal AI agent designed to go beyond answering questions and actually perform tasks across users’ digital lives, including sending emails, booking travel, making purchases and managing information across connected applications.
Initially available in the US through a dedicated app and WhatsApp, Muse is the centrepiece of CEO Mark Zuckerberg’s push towards what Meta calls “personal superintelligence”.
The company says the agent can continue working in the background after users close the app and return when a task requires approval.
AI That Can Act on a User’s Behalf
Muse can connect to applications covering email, calendars, shopping, health, payments and smart-home services. Users decide which services the agent can access and can revoke permissions at any time.
Meta has built Muse around a dedicated virtual machine called Muse Secure VM, which keeps the agent and a user’s data in an isolated environment. A separate Sentinel system monitors actions, with Meta saying it can require user approval before sensitive activities such as sending an email or making a purchase.
The agent is free for basic use, while Meta is offering subscription plans priced at US$20 and US$100 per month for heavier usage, according to the company. Meta also plans to bring Muse to its AI glasses.
Safety Concerns Shadow the Launch
The technology arrives with significant safety challenges. Meta delayed Muse’s release earlier this year while it worked on security measures, according to company executives.
Internal testing has nevertheless produced mixed results. Employees have reported reliability problems and cases where the system allegedly bypassed safeguards to expose personal iCloud photos. Other testers reported problems with background monitoring and repeated logouts. Meta did not comment on the specific incidents reported internally.
The concerns highlight a broader challenge for AI agents. An assistant that only provides information can make a wrong recommendation. An agent with access to email, payments and personal files can potentially turn that mistake into a real-world action.
Meta is betting that its security architecture can make such agents safe enough for mainstream use. The company says it will introduce an encrypted version of Muse later this year, with the encryption key controlled by the user.
Source: Reuters
Vishal is an experienced Editor at Inspirepreneur Magazine with key interests in artificial intelligence, eCommerce, entrepreneurship, lifestyle and startup sector. Prior to joining Inspirepreneur, he was a Content Writer cum Correspondent at Siliconindia Magazine, where he worked on Company Profiles, Cover Stories, Executive Profiles, Feature Articles and Thought Leadership content.
You Might Also Like
Sweden’s EQT closes record $15.6B Asia Fund, region’s biggest ever
200 Dead in Philippines as Typhoon Kalmaegi Devastates Vietnam
