Meta introduced Muse on September 8, 2026, a personal AI agent designed to execute tasks rather than merely answer prompts. The agent can manage email, travel tickets, retail shopping and payments, and is initially available only in the U.S. across the web, iOS and Android, with Meta's AI glasses to follow later.
Muse can open a browser, complete forms, negotiate deals and purchase products. For payments, it uses Stripe's Link infrastructure, which creates a one-time card number so the agent never sees the user's actual payment card. Meta says additional payment options, including Shop Pay and 1Password, are expected soon. Muse has access to services such as Gmail, Spotify, Ticketmaster, OpenTable and Shopify. It is powered by the Muse Spark family, launched on April 8 and updated to Muse Spark 1.3 on September 2.
Meta is leaning on distribution through WhatsApp, which has more than three billion users, while offering a free tier and paid subscriptions at $20 and $100 per month. Security is a central selling point: Muse operates in a dedicated Muse Secure VM, with a separate Sentinel system acting as permission authority. Meta says sensitive actions can require explicit approval, and its bug bounty may reward up to $300,000 for vulnerabilities. The company also plans a Muse Confidential VM later this year, encrypting the environment with a user-controlled key.
The launch comes 13 days after Meta agreed to settle lawsuits involving Facebook and Instagram for up to $18 billion over potential child harm, heightening the trust challenge around handing an AI access to inboxes and credit accounts. McKinsey's 2026 AI Trust Maturity Survey found that two-thirds of respondents named security and risk as the main obstacle to implementing agentic AI.
On the same day, U.S. officials accused six China-based AI firms including DeepSeek, Moonshot AI and Alibaba of industrial-scale distillation using outputs from leading American AI models. The NSA, FBI and CISA issued a joint advisory saying the companies targeted proprietary capabilities from Anthropic, OpenAI and Google's AI operations. Washington claimed the activity likely had Chinese government awareness and could advance China's military and cyberattack capabilities, though Beijing had no immediate response.
Together, the developments highlight an AI race moving beyond chatbot responses and into payments, identity, security and real-world transactions. Gartner forecasts stand-alone AI agents will grow 13-fold from 2025 to 2030 into a $222 billion opportunity, while BCG estimates agentic AI could create up to $200 billion in net new technology-services demand over five years.
