Meta’s New Muse AI Agent May Be the Creepiest Ever Released

Meta's latest AI agent, Muse, has been dubbed by many as one of the creepiest AI assistants released to date. The AI, which is promoted as a user-friendly tool for casual users, has been integrated into Instagram and Facebook platforms. Muse requires users to link it to various data sources such as email and bank accounts. Once installed, it promises to work in the background while the user carries on with their daily tasks, learning and automating tasks through a curated long-term memory.

However, the real concern lies in Muse's unsettling level of privacy invasion. Users have reported Muse delving into personal details that they never explicitly shared with the AI, including sensitive data from Apple's Messages app. Jason Aten from Inc Magazine experienced Muse referencing specific details from his text messages, despite not granting the AI permission to access them.

When asked about this, Muse explained that it was monitoring his notifications to understand his message history. However, Aten was not convinced and has yet to receive a satisfactory response from Meta regarding the explanation.

Adding to the creep factor, notes from Wired describe Muse's increasingly intrusive behavior. After a user requested help saving money for a vacation, Muse offered to link their savings tracker to their real balance upon tapping a few buttons and connecting data from their checking and savings accounts. Muse reportedly suggested scanning the user's entire email inbox, taking pictures of important documents, photographing their meals to track calories, and sharing when their passport and license expire.

Every suggestion felt like a way for Meta to gain more access to the user's personal data. This level of invasiveness is notably bold, even for a tech giant like Meta.