Meta launches Muse AI agent weeks after $18 billion settlement, betting on trust

Meta launches Muse, its biggest consumer AI bet

Meta has launched Muse, a personal AI agent designed to carry out everyday tasks such as shopping, sending email and planning travel. The assistant is rolling out in the US on iOS, Android and the web, and can also be reached through WhatsApp, with Meta's AI glasses to follow. The launch lands less than two weeks after Meta agreed to an $18 billion multistate settlement over social media's harms to consumers, according to TechCrunch — awkward timing for a product that asks people to hand over more personal information than the company has ever requested before.

What Muse does

Muse is built to act rather than just chat. Once given a goal, it can open a browser, fill out forms and even negotiate on a user's behalf, The Verge reports. For longer jobs it keeps working in the background after the app is closed, returning only when something changes or when it needs approval, such as for a purchase.

Under the hood, Muse connects to the apps and services people already use — email, calendars, payments, health and fitness, smart home, dining and events, according to TechCrunch. It ships with built-in connectors, can wire itself to services that expose a public API using credentials the user supplies, and falls back to plain browser access when neither option exists. Checkout runs through Stripe's Link, with Shopify's Shop Pay and 1Password support planned.

The agent is powered by Meta's in-house Muse Spark model and remembers details users mention even once, enabling unprompted suggestions. Users can customize it with a name, an avatar and communication settings, TechCrunch notes. Rival agents already exist — The Verge points to OpenAI's ChatGPT Work, Anthropic's Claude Cowork, Microsoft's Copilot Tasks and Google's Gemini Spark — but Meta argues Muse stands out on ease of use, with no learning curve and a messaging-style interface aimed at non-technical users.

Pricing and availability

Muse is free for most usage, though it requires a payment card at signup and subscriptions kick in as usage grows. TechCrunch reports two paid tiers at launch: Power at $20 per month and Maximum at $100 per month, alongside a usage meter that warns people before charges apply. Meta believes most users will stay on the free tier, TechCrunch adds, while The Verge notes the announcement did not spell out free-tier limits.

The security pitch

Because agents hold credentials and payment details, Meta has built Muse to run on a dedicated cloud virtual machine with its own browser, isolated from other users' agents. A separate watchdog agent, Sentinel, runs on the same machine but is kept apart from Muse at the system level, and nothing Muse does reaches the internet without approval. Meta says Muse has no visibility into passwords or payment methods and that conversations are not shared with its advertising systems.

Users can opt out of their interactions being used to train Meta's models — The Verge notes it is unclear whether that setting is on by default — and can tell Muse to forget specific things it has learned. A more secure "confidential" virtual machine, encrypted so that even Meta cannot access it, is planned for later this year. TechCrunch cautions that Meta's security claims, laid out in a technical post, still need scrutiny from outside security experts.

Meta's trust deficit

Meta's history works against it. Beyond the August settlement with 29 states, the company was ordered to pay $942 million in a New Mexico case over harms to children, and thousands of personal-injury and school-district cases remain pending, TechCrunch reports. Its record includes a then-record $5 billion FTC privacy settlement in 2019, a 2023 FTC charge that it violated the resulting order, and a 2019 disclosure that user passwords had been stored in readable form. Cambridge Analytica still looms, The Verge adds, alongside recent AI missteps: a Discover feature that exposed other users' prompts, a support chatbot that helped hackers take over more than 20,000 Instagram accounts, and a short-lived AI-generated clickbait news feed. The wider category faces the same headwind — TechCrunch notes that early testers of the assistant Instinct were startled by license terms granting broad rights over user material.

Why it matters

Muse is Meta's biggest consumer AI bet to date and the centerpiece of a strategy to catch OpenAI, Anthropic and Google after years of setbacks. But agents demand deeper trust than social media ever did: they need logins, inboxes and wallets. With billions of people reachable through apps like WhatsApp, Meta can put an agent in front of more casual users than any rival — which makes whether consumers forgive its track record a consequential question for the entire consumer AI category, not just for Meta.