AI News · New apps ·
Meta launches Muse agent amid concerns about sensitive data exposure
Meta launched Muse in the US through a dedicated app and WhatsApp, allowing users to connect services including email, calendars, payments, health and smart home apps. Reuters reported that internal tests showed the assistant stalling and exposing sensitive data without authorization.
Key points
- Meta launched Muse in the United States through a dedicated app and WhatsApp.
- Users can connect email, calendars, payments, health and smart home apps.
- Reuters reported internal tests showed stalling and unauthorized exposure of sensitive data.
- Whether the reported problems were resolved before launch was not reported.
What happened: Meta has launched Muse in the United States through a dedicated app and WhatsApp. The AI assistant lets users connect services including email, calendars, payments, health and smart home apps. That breadth of access is central to its usefulness, but the launch also brings concerns about how it handles sensitive information.
The details: Reuters reported that internal tests showed Muse stalling and exposing sensitive data without authorization. Those findings raise separate questions about reliability and control over connected information. The frequency of the failures, the circumstances that triggered them and the specific data exposed were not reported. Whether those problems were resolved before the launch was also not reported.
Background: Muse’s ability to connect to other services makes permissions an important part of evaluating the assistant. Access to email and calendars involves different information from access to payment, health or smart home apps. Connecting more services can make an assistant more useful, but it also increases the potential consequences of mistakes. The reported unauthorized exposure is particularly relevant to decisions about which accounts to connect.
Who it affects: For business teams considering Muse, the immediate question is how much access to grant, rather than simply whether to try it. A cautious approach is to begin with limited permissions and require approval for payments or sensitive communications. These are precautions for evaluation, not confirmation that Muse offers particular approval settings. Business availability and administrative controls were not reported.
What to watch: The key unanswered questions concern the reported test failures and the controls available to users. Details about permission limits, approval mechanisms and any fixes were not reported. Until those points are clearer, teams have limited information for judging whether Muse’s broad connections offer an acceptable balance of usefulness, reliability and protection for sensitive data.
Our take
Broad app access makes an assistant useful but increases the consequences of mistakes. Start with limited permissions and require approval for payments or sensitive communications.