🔍 Read the full analysis: Meta’s AI Agent May Read Messages Beyond Muse: What To Know on ThorstenMeyerAI.com
Get the latest gadgets delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
TL;DR
Meta’s Muse personal AI agent can use connected services, including Meta apps, email and—if enabled—Mac Messages. Reporting by WIRED and TIME says its instructions direct it to maintain regularly updated pages about people mentioned in a user’s communications, including people who never signed up. Meta says each user controls Muse’s access; the exact reach of the profiling and deletion limits remain important open questions.
Meta’s Muse personal AI agent can access communications through connected services, and reports by WIRED and TIME say its instructions tell it to maintain regularly updated pages about people in its user’s life—including people who never signed up. Meta has not disputed the reported instructions and says each user controls what Muse can access, leaving questions about how people mentioned in those communications can limit the resulting information about themselves.
Muse uses connectors to reach services on a user’s behalf. Meta’s help materials say email connections let Muse search messages but do not download an entire inbox unless instructed. Calendar connectors may send updates proactively. Facebook, Instagram and Threads can connect automatically when accounts are in the same Accounts Center. Meta chief AI officer Alexandr Wang said the Instagram connection includes direct messages and that Messenger is also available; Meta’s help pages do not specify the full scope of those connectors.
Access to Mac Messages requires the user to grant macOS Full Disk Access and enable the connector. Meta says Muse runs in a private virtual machine assigned to each user. In WhatsApp, Muse is a chat a user opens; the source material does not say that opening it gives the agent access to the person’s other WhatsApp chats. Custom connectors are another route, though Meta’s help page says it does not review them.
WIRED reported on Muse instructions obtained by security researcher Karan Joshi, and TIME later reviewed instructions separately. The reports say the agent is directed to create a page for each person in the user’s life and refresh information hourly, drawing on chats, messages and emails it has read. The pages may include details about relationships, shared interests and disagreements. Meta did not dispute TIME’s findings. The report’s account of what Muse is instructed to do is distinct from a complete public explanation of when or how each connector supplies data.
Infographic 01 · What Muse can read
Muse reads what its user can read. That includes you.
Connectors decide what the agent sees. The user switches them on. The people in that user’s chats and inbox are never asked.
Sources a Muse user can open to the agent
The agent reads what its user can read. Everyone who writes to that user is inside.
| Source | How it connects | Status |
|---|---|---|
| Email (Gmail, Outlook) | Opt-in connector. Searches when a task needs it; downloads the whole inbox only if told to. | Meta document |
| Calendar | Opt-in connector. Can push changes to Muse on its own (“proactive”). | Meta document |
| Facebook, Instagram, Threads | Connected automatically if the accounts share an Accounts Center with Muse. | Meta document |
| Instagram DMs, Messenger | Named as Muse-only connectors (“Instagram (incl dm’s) … and Messenger”). What they read is not documented. | executive post |
| Mac Messages (iMessage) | Only with macOS Full Disk Access plus the Messages connector, Meta says. One columnist says it synced anyway. | disputed case |
| A place to chat with Muse, not a connector. Test-build text mentions Muse access to end-to-end encrypted chats. | leaked code | |
| Custom connectors | Muse can build links to other services. Meta does not review them. | Meta document |
What the instructions say
Extracted by researchers; reported by WIRED and TIME reported
- “A page for every person in the user’s life”: family, partners, friends, colleagues, people followed
- Notes on how people met, shared interests, disputes, “tensions and alliances”
- People who never signed up are mapped by other users’ agents
- De-identified “lessons” shared across agents; Meta did not dispute TIME’s findings
What Meta says
Privacy policy, help pages, spokespeople Meta document
- Each user’s VM is isolated; no other user’s agent can reach it
- Conversations and VM data are not shared with Meta’s ad systems
- “Each person stays in control of their Muse and decides how much access it gets”
- Forgetting works “to the best of its ability”; deleted data may still be remembered
Sources: Meta Help Center: Connectors; Muse Privacy Policy; Alexandr Wang on X (8 Sep 2026); TIME (6 Oct 2026); Tom’s Guide on WIRED’s report (6 Oct 2026); TechCrunch (30 Sep 2026); Digit (16 Oct 2026). Status labels: Meta document = stated by Meta; executive post = Meta executive on X, not in help pages; disputed = claim and denial on record; leaked = unreleased code, may never ship.
Powered by Thorsten Meyer AI
When Other People’s Messages Feed Muse
The issue extends beyond people who choose to use an AI agent. If a Muse user grants access to a conversation, information about the other participants may also be processed and organized. That creates a privacy gap between the person who grants access and the people discussed: the account holder has settings, while a contact may have no direct relationship with Muse or its controls.
For readers, the practical question is not only whether they installed Muse, but what people they communicate with have connected. Group chats and email threads can include personal details about several people. The reporting does not establish that every message is continuously read or that every person mentioned receives a profile; it does describe instructions to keep pages on people named in accessible communications.
The commercial context may also matter. Meta announced a large stock sale alongside an investment from Nvidia, whose chips Meta plans to deploy for Muse. That financing does not prove that the agent’s business model depends on broader personal data access. But the expansion makes the company’s choices about connectors, retention and user controls relevant to a growing audience.
Popular right now

Bose QuietComfort Wireless Bluetooth Headphones with Active Noise Cancellation, Mic, USB‑C Charging, Deep Bass, 24‑Hour Battery – White Smok

De'Longhi Magnifica Evo Fully Automatic Espresso Machine with Manual Milk Frother

Roborock Qrevo CurvX Robot Vacuum and Mop
As an affiliate, we earn on qualifying purchases.
Muse’s Rollout and Data Connections
Meta launched Muse in the United States on September 8, 2026, describing it as a personal AI agent that can act for users, such as handling email or making bookings. Japan followed on October 23, and Meta announced Singapore availability on October 27. TIME reported more than five million downloads within weeks; that figure is TIME’s report, and the source material does not provide a measurement method.
The privacy questions became public through reporting in October. WIRED published findings based on files Joshi said he obtained by asking Muse in chat. TIME reviewed the instructions separately three days later and described hourly updates to information about the user and people mentioned in communications. Meta’s policy says Muse may collect messages, connector data and information about the user or others. These reports concern the agent’s documented instructions and data practices, not independent proof that every described profile is complete or accurate.
Meta’s help material also distinguishes forgetting from deletion. It says Muse may still remember information learned from something a user deleted, and describes forgetting as working “to the best of its ability.” The source account says users can disconnect connectors, turn off model training, edit a memory file or reset Muse. Those options apply to the Muse user’s account; they do not give a person mentioned in someone else’s messages direct control over that user’s archive.
“Muse has connectors like Instagram (incl dm’s), Threads, Facebook (incl marketplace), and Messenger.”
— Meta chief AI officer Alexandr Wang, writing on X
The Limits of Muse’s Memory
Several details remain unclear. Meta’s public help pages do not fully explain what Instagram and Messenger connectors retrieve, how often those services share information, or precisely when Muse accesses messages. The reports describe instructions to refresh pages hourly, but the material supplied does not independently establish how consistently the system carries them out or how accurate its records are.
It is also unclear what information remains after a user disconnects a service, deletes a conversation or asks Muse to forget something. Meta says forgetting may not remove information already learned, and the reported instructions say the agent should not point out that original messages may remain visible. The source material does not provide a detailed retention schedule or explain whether people who did not sign up can request access to, correction of or deletion of a page created from another person’s communications.
The financing figures are reported as a $25 billion stock sale at $625 a share, plus $12.5 billion of preferred stock, with Nvidia investing $3 billion. The source material does not establish the terms or privacy conditions of Nvidia’s investment, or show that the financing changes Muse’s access to user data.
Watch for Controls and Policy Detail
The immediate next steps are Meta’s further explanations of connector access, memory retention and the options available to people who appear in another user’s Muse data. Readers using Muse can review connected services, disconnect ones they do not need, check model-training settings, inspect or edit the memory file, and reset the agent. These settings can change what that user’s Muse can access; they do not necessarily remove material already visible in other people’s accounts.
People who do not use Muse have fewer direct options. The source material points to WhatsApp’s Advanced Chat Privacy as one way to limit what can be moved out of a conversation, but it does not establish a universal control across email, Instagram, Messenger or other services. Meta’s expansion into Singapore and its existing rollout markets may bring further clarification as availability grows. Until Meta publishes more detail, the exact scope of processing and the remedies for nonusers remain unresolved.
Key Questions
Does Muse read every message in every connected inbox?
Not according to Meta’s help materials cited here. They say email connection allows searching but does not download an entire inbox unless the user instructs it to. The source material does not establish that Muse reads every message across all connected services.
Can someone be included in Muse’s records without signing up?
WIRED and TIME reported that Muse instructions direct it to maintain pages about people mentioned in a user’s accessible chats, messages and emails. That means a person may be described in information available to another user’s agent without having a Muse account. The reports do not establish that every person mentioned is profiled.
Does Muse have access to all WhatsApp chats?
The source material says Muse in WhatsApp is a chat a user opens, not a reader of that user’s other chats. It also mentions Advanced Chat Privacy as a way to limit what can be taken out of a conversation.
What can Muse users do to limit access?
Users can review or disconnect connectors, switch off model training, edit Muse’s memory file or reset the agent, according to the source material. Meta says forgetting works “to the best of its ability,” so asking it to forget something should not be treated as a guarantee that all learned information is deleted.
Can nonusers delete information Muse learned about them?
The source material does not identify a direct control for nonusers to inspect or delete information in another person’s Muse archive. Meta has said the Muse account holder controls the agent, but the scope of options for people mentioned in that account remains unclear.
Source: ThorstenMeyerAI.com
