Meta Debuts Muse, Its Long-Planned Personal AI Agent

Meta has unveiled Muse, a personal AI agent created under chief AI officer Alexandr Wang. The product, a long-time development, was highlighted as a major step by CEO Mark Zuckerberg in his recent extensive statement, according to Axios. Muse operates within a chat interface, akin to a text conversation. It has been engineered to be more proactive and persistent than regular chatbots.

Users can assign a name to their agent, generate an avatar, and personalize its communication style. The Muse agent operates on a dedicated virtual machine in Meta's cloud, featuring a user-visible built-in browser. Meta provides a free version of Muse, along with two subscription options at $20 and $100 per month. For the majority of users, the free tier should suffice, but the subscriptions aid Meta in covering computational expenses.

There is no advertising within Muse, though Wang hinted at exploring commerce opportunities that could generate extra income. Muse initially launches in the U.S. for iOS, Android, and the web, with plans for integration with Meta's AI glasses. Wang envisions the long-term goal of developing personal superintelligence to assist individuals in achieving their objectives, fulfilling their passions, and creating things beyond their reach without technological aid.

Meta offers enhanced privacy controls with Muse compared to its prior AI offerings, including the ability to prevent queries from being utilized by Meta and a confidential mode where the company cannot view activity in a user's virtual workspace. Additionally, there is a separate system called Sentinel that oversees Muse's internet and connected service access.