Meta introduces personal AI agents amid scrutiny of privacy and safety

What This is

• Meta launched Muse, a personal AI agent that can carry out tasks on a user’s behalf, such as sending emails, booking travel, shopping, and filling out online forms.
• Muse can connect to user-approved apps across areas like email, calendars, payments, shopping, health, and smart-home services, and users can revoke access.
• A health-data company, HealthEx, says it has built a connector that lets users link medical records to Muse if they choose, so the agent can help prepare visit summaries and reminders.
• Muse is offered with a free tier and paid subscription plans for higher use limits.
• Meta says each Muse agent runs in a separate cloud “virtual machine,” meant to isolate the agent’s activity and personal data from other systems.
• The launch comes as Meta faces legal and public scrutiny over privacy and safety, including large settlements and new requirements aimed at protecting young users.

Why This Matters

• When an AI can take actions in your accounts, errors or abuse can turn into real-world harm, like unwanted payments, messages, or exposure of sensitive records.
• These tools shift power and responsibility from users to software, raising new questions about consent, accountability, and what “personal control” means in a digital life.
• Meta’s existing trust and youth-safety problems make it a test case for whether big platforms can persuade people to hand over broad access to an AI agent.