Full Breakdown
Meta Launches Muse, Its First Consumer AI Agent Amid Trust Concerns
9/9/2026, 12:34:38 AM
Launch of Muse Personal AI Agent
On Tuesday, Meta unveiled Muse, a personal AI agent for U.S. adults. The service is available via a Muse app on iOS and Android, the website muse.ai, and WhatsApp. Users can ask Muse to send emails, book travel, shop, manage calendars, and pursue longer-term projects such as fitness plans or business setups. The agent can continue working after the app is closed and will return with updates or when user approval is needed.
Background & Context
Muse is part of Meta’s multi-billion-dollar AI overhaul led by Meta Superintelligence Labs, created by CEO Mark Zuckerberg last year. The launch follows Meta’s settlement of an $18 billion multistate lawsuit over social-media harms and earlier FTC penalties. Meta positions Muse as a way to catch up with rivals such as OpenAI, Anthropic, and Google, which already offer agentic tools for developers and enterprises.
Data & Statistics
- Two subscription tiers at launch: Power at $20/month and Maximum at $100/month.
- Free users receive up to 100 million tokens per week; an in-app meter warns when the allowance is exhausted.
- Muse connects to Gmail, Google Calendar, Spotify, Ticketmaster, Shopify, Stripe Link, and upcoming integrations with 1Password and Shop Pay (The Verge, TechCrunch).
- The agent runs inside a dedicated “Muse Secure VM” that isolates user data from other agents and from Meta’s central infrastructure.
Official Statements & Responses
Meta notes that users can opt out of having their interactions used to train its models and can command Muse to forget specific information.
Criticism & Opposition
Consumer-trust experts point to the contrast between Muse’s promised privacy safeguards and Meta’s history of privacy violations, citing the 2011 FTC settlement, the $5 billion 2019 penalty, and a 2023 FTC order. TechCrunch highlighted that earlier AI assistants required “perpetual and irrevocable” licenses to use user data, raising concerns that Muse could repeat similar overreach. CNBC reported that, despite Meta’s technical assurances—such as a separate Sentinel agent that approves sensitive actions—skeptics question whether policy-based protections are sufficient without cryptographic guarantees.
Why It Matters
Muse marks Meta’s shift from a social-media-centric model to a broader AI-driven revenue stream, directly targeting everyday consumer workflows. Success would show that a company with a troubled privacy record can regain user confidence while competing in the emerging market for autonomous AI agents.
What’s Next
Meta plans to release a “confidential” version of the Muse Secure VM later this year, promising encryption with a user-held key. The company is also expanding its bug-bounty program, offering rewards up to $300,000 for vulnerabilities that affect Muse. Subscription pricing and feature sets may evolve as the service moves from limited rollout to broader availability.
