More security and control

Meta Launches Personal AI Agent “Muse”

Meta AI
Facebook
X
LinkedIn
Reddit
WhatsApp
Source: QubixStudio/Shutterstock.com

Meta releases a personal AI agent with “Muse”. The system acts proactively in its own virtual environment.

Meta has officially presented its new personal AI agent “Muse”. Developed by Chief AI Officer Alexandr Wang, the system is based on Meta’s long-term strategy for creating intelligent digital assistants, which CEO Mark Zuckerberg recently highlighted in a manifesto.

Ad

Features and system architecture

  • Proactive interaction: Muse operates via a chat interface, works long-running as well as proactively, and can be customized individually in name, avatar, and communication style.
  • Isolated work environment: The agent runs on a dedicated virtual machine in Meta’s cloud and uses an integrated browser visible to the user.
  • Platforms: The service initially launches in the US for iOS, Android, and the web; a direct connection to Meta’s AI glasses is expected to follow shortly.

Pricing structure and business model

Meta provides a free basic version, which is intended to be sufficient for the majority of users. For computationally intensive applications (“power users”), the company offers two paid subscriptions for 20 and 100 US dollars per month, respectively, to cover computing costs. Currently, no ads are displayed within Muse, but Meta is examining future e-commerce opportunities.

Enhanced data privacy and safety system “Sentinel”

Compared to previous AI features, Meta equips the system with additional security and control mechanisms:

  • Opt-out for training data: Users can prevent their prompts from being used to train future models.
  • Confidential mode: A planned Confidential Mode is intended to ensure that Meta has no insight into activities in the virtual workspace.
  • Security instance Sentinel: The interface between the agent, the internet, and connected services is monitored by a separate security system called “Sentinel”.

“The vision for the future is the development of a personal superintelligence that helps people achieve their goals, pursue their passions, and create things that would never have been created without this technology.”

Ad

Alexandr Wang, Chief AI Officer at Meta

(Editorial Team)

Ad

Weitere Artikel