Meta launches AI agent capable of accessing other apps

Meta has launched a personal artificial intelligence agent capable of accessing other applications to send emails, arrange travel, and make payments, taking the company deeper into technology that can act autonomously on behalf of users.

Called Muse, the service is initially available in the US through a dedicated app and WhatsApp. Users can connect it to services including email, calendars, payments, shopping, health applications, and smart-home systems, and can revoke those permissions.

Tap here to add The Daily Star as a trusted source

Meta says Muse runs in a dedicated virtual machine and includes safeguards designed to monitor proposed actions and request additional authorisation for sensitive tasks. A free version is available alongside paid plans costing $20 and $100 a month for heavier use.

The launch nevertheless comes with significant security questions. Reuters reported that Meta employees conducting internal tests had encountered reliability problems and incidents involving unintended exposure of sensitive information, including private photographs.

Meta said it delayed an earlier launch to strengthen the system and believes it has now reached the minimum safety standard required for release.

Muse is central to chief executive Mark Zuckerberg’s effort to turn Meta’s heavy spending on AI infrastructure into consumer products and build what he has described as personal superintelligence.