Meta has launched Muse, a personal AI agent designed to handle everyday tasks rather than simply answer questions. Initially available in the US, Muse can browse the web, connect with apps and services, book appointments, complete forms and continue working in the background.
Muse can operate through its dedicated app and WhatsApp, while Meta has also introduced versions for iPhone, Android and Mac. Users can connect services such as email, calendars and Instagram, with permissions controlled by the user.
The technology uses a dedicated virtual machine with its own browser. Meta says this allows Muse to perform multi-step tasks while keeping the agent separated from a user's main computing environment. Critical actions, including purchases and certain messages, require user approval.
Privacy and security have been central to the rollout. Meta says login credentials are stored separately from the agent and that users can review activity, manage permissions and revoke connections.
The launch follows earlier scrutiny of Meta's AI image-generation products and comes as major technology companies compete to develop increasingly autonomous AI agents. Reuters has also reported that Meta identified security and reliability issues during Muse's development before its public launch.
0 comments
No comments yet. Yours would be the first.