Meta’s AI Muse: Reports on profiles of friends and family

US tech giant Meta is once again facing scrutiny from privacy advocates over its AI assistant, Muse. According to a report by Golem.de, the AI agent compiles profiles on users’ friends and family members. As a result, the collection of personal data affects not only users themselves, but also people in their immediate circles who may not even use the service.

Profiling of Social Circles

According to reports, Meta’s AI assistant analyzes the interactions and information that users share in everyday life, using it to build structured profiles of contacts. When it was first introduced, the company presented the technology as the personal AI agent Muse, designed to handle tasks autonomously. More recently, Meta expanded the technology’s capabilities to allow autonomous phone calls by AI agents. However, the fact that these systems also gather and catalog information on uninvolved third parties marks a delicate expansion of data collection.

The profiling is rooted in common digital platform mechanisms: gathering user traits typically serves to personalize content and build target audiences. When such profiles are inferred and stored, it threatens to create a detailed mapping of social relationships without the explicit consent of those affected.

See also  Synology fixes critical Pwn2Own vulnerabilities

Data Protection Concerns

From a data protection perspective, this practice is highly problematic. When users share details from their personal lives with an AI assistant—such as birthdays, preferences, addresses, or relationship dynamics—this information must not be processed into standalone profiles without a legal basis.

Security concerns surrounding the software have already surfaced in the past, such as when a vulnerability in Meta Muse for macOS came to light. The creation of shadow profiles covering users’ personal circles further fuels the debate over the transparency of AI agents.

What users should consider now

Anyone using AI agents like Muse in their daily routine should be aware of the implications. Users should avoid entrusting the assistant with sensitive details about relatives, colleagues, or friends. It is also advisable to carefully review the Meta account’s privacy settings and restrict the assistant’s access to contacts or chat histories to the absolute bare minimum.

Sources: Golem.de/security

Leave a Comment

Your email address will not be published. Required fields are marked *

Mastodon
Scroll to Top