DALLAS - TelAve -- DALLAS, Texas -- A comprehensive new report released today by SaNBi dives deep into the growing controversy surrounding Meta Muse, the newly launched personal AI agent. In its first few weeks of operation, the AI agent has been at the center of multiple privacy and security alarms, raising serious questions about the safety of handing over standing authority to artificial intelligence.

The detailed breakdown highlights a series of alarming incidents. Among the most concerning claims is that the Muse Mac app read a user's private iMessages without explicit permission. While Meta has disputed this claim by stating that access is strictly opt-in, the AI agent itself provided an incorrect explanation of how it accessed the data, sparking widespread concern over AI permissions.

In another shocking case, the AI agent reportedly shared a Facebook Marketplace seller's home address with a buyer under an "Allow Always" setting. The agent also incorrectly informed the buyer that the seller was home. Meta has stated that no privacy controls were technically breached, highlighting the dangerous complexities of agentic AI permissions where a single authorization can lead to severe real-world consequences.

More on TelAve News

• NIL Club Data Shows Brands Are Looking Beyond Follower Counts in College Sports
• SafecomLink 5G-AI-ALE: Automatic Link Establishment that Stays Active for the Life of the Link
• Georgia's Magical Nights of Lights Bigger and Brighter for 2026
• Harvard Educated Board Certified Endodontist Joins Pearlfection Dentistry
• NYC Banners Delivers High-Stakes UN General Assembly Backdrop for the US Department of State Event

Adding to the controversy, security researchers have already discovered a Mac app zero-day vulnerability and a cloud flaw within the first three weeks of the product's launch. Furthermore, major retailers like Amazon have reportedly blocked Meta Muse from their platforms, signaling a brewing battle over automated agent access and security.

SaNBi notes that the current pattern is about permission rather than just hacking. Most Meta Muse privacy incidents turn on what users technically allowed versus what they believed they allowed. While a standard chatbot error is just a bad answer, an AI agent error can result in a sent message, a shared physical address, or an unauthorized purchase.

To read the full timeline of the Meta Muse controversy and find actionable safety advice for users and security teams, visit the full report here: https://sanbi.ai/blog/meta-muse-controversy-privacy-security

As AI evolves from simply generating text to taking action across accounts, experts urge users to exercise caution and keep agent permissions as narrow as possible.