🔍 Read the full analysis: Can Avoiding Muse Limit Meta’s AI Agent’s Access To Messages? on ThorstenMeyerAI.com
Get business pricing on office and shipping supplies
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
Meta’s Muse personal AI agent can search information available through a user’s connected accounts, including some messages, and Meta’s policy says it may collect information about other people. WIRED and TIME reported that Muse instructions direct it to maintain profiles of people mentioned in users’ communications; Meta did not dispute the findings and says users control access. People who do not use Muse have limited direct controls over information held in someone else’s account.
Meta’s Muse AI agent can access some messages and other account data when a user connects services, while reports say its instructions direct it to maintain profiles of people mentioned in those communications. The findings matter to people who have never used Muse: information about them may still be processed through another person’s account, though the precise scope and retention of that information are not fully clear.
Muse uses connectors to interact with outside services. Meta’s help information says an email connection lets the agent search messages, but does not download an entire inbox unless the user asks. A calendar connector can send updates to Muse. Meta says Facebook, Instagram and Threads accounts may connect automatically when they are grouped in the same Accounts Center. The company’s chief AI officer, Alexandr Wang, said the Instagram connector includes direct messages and that Messenger is also supported; Meta’s public help pages do not fully document what those connectors can access.
Access to Apple’s Mac Messages requires the user to grant macOS Full Disk Access and enable the connector, according to Meta. Muse can also use custom connectors, which Meta says it does not review. Meta describes its agent as operating in a private virtual machine for each user, and says Muse in WhatsApp is a chat a person opens—not a tool that automatically reads their other WhatsApp conversations.
WIRED and TIME separately examined Muse instructions that reportedly tell the agent to create and update pages about people in a user’s life, drawing on names and details in chats, messages and emails. TIME reported that the instructions call for hourly updates. Meta did not dispute the reports’ findings. A company spokesperson said: “Each person stays in control of their Muse and decides how much access it gets.” That control applies to the Muse user; it does not give people mentioned in the user’s messages the same settings.
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
How Other People’s Messages Are Exposed
The issue is that consent to use an agent is not the same as consent from everyone in a conversation. A person who connects an inbox or social account may authorize Muse to process material that includes family members, colleagues or friends. Those contacts may not know their messages are available to the agent, and they may have no Muse account through which to manage the information.
This matters because personal communications can reveal relationships, disputes, plans and other details beyond the immediate task a user gives an AI. The reports describe instructions to build profiles and refresh them, while Meta’s policy says it can collect information about “you or others.” The precise extent to which those profiles are generated, retained or used in particular cases remains uncertain. Still, the documented access model puts responsibility mainly on the account holder, even when the data involves several people.
For users, the practical question is not only whether Muse can perform a task, but what account access that task requires and whether the same result can be achieved without connecting sensitive communications. For non-users, the available options are more limited: they can ask contacts to restrict sharing or change settings in particular messaging services, but cannot directly control another person’s Muse configuration.
Products Worth Considering
Dell XPS 13 9350 Laptop, 13.4 FHD+ Display, Intel Ultra 7 256V, 16GB DDR5, 1TB SSD
Brother TD-2020A 2‑Inch Direct Thermal Desktop Label Printer with USB
Logitech MX Brio Ultra HD 4K Webcam, Graphite
As an affiliate, we earn on qualifying purchases.
Muse’s Connectors and Rollout
Meta introduced Muse in the United States on September 8, 2026, describing it as a personal AI agent that can act on a user’s behalf, including handling email, bookings and shopping. Japan became its first market outside North America on October 23, and Meta announced availability in Singapore on October 27. TIME reported that Muse had passed five million downloads within weeks; that figure is TIME’s report, not a measure of how many people enabled each connector.
Meta’s system relies on permissions and connections rather than one uniform level of access. Some connectors require explicit setup, while the company says certain Meta accounts in a shared Accounts Center connect automatically. Its privacy materials also describe connector data that may be shared proactively. The distinction between what an agent can access, what it retrieves for a particular task and what it stores afterward is central to understanding the reports.
The source report also linked Muse’s expansion to Meta’s financing and chip supply. It said Meta sold $25 billion in stock at $625 a share and issued $12.5 billion in preferred stock, while Nvidia was investing $3 billion as Meta expands its work on Muse. Those figures do not establish how the financing will affect privacy practices, but they underline the commercial scale of the agent’s rollout and infrastructure needs.
““They’re trying to know you like a friend, which is honestly pretty creepy.””
— Karan Joshi, security researcher, speaking to WIRED
What Muse Retains After Access
Several operational details remain unresolved. Meta’s public materials do not specify exactly what data its Instagram and Messenger connectors retrieve, how long information about non-users remains in Muse’s files, or whether every reported profiling instruction is applied in the same way across users and tasks. The reports describe instructions and system behavior, but do not establish that every person mentioned in a message receives a complete or accurate profile.
Meta’s help material says that after a user deletes information, Muse may still remember details it learned from it and that it will forget information “to the best of its ability.” That wording leaves open what deletion removes from underlying messages, derived memories and other records. The source report also says TIME found instructions not to point out that original messages may remain visible; Meta’s response did not publicly address that specific point.
It is also unclear what safeguards apply to information shared as de-identified lessons across the system, as described by TIME. De-identification does not by itself answer how those lessons are selected, retained or checked against re-identification risks. Meta’s statement that each user controls access does not resolve these questions for people who do not use the product.
Settings Users Can Change Now
Muse users can review connected services, disconnect connectors they do not need, adjust the model-training setting, edit the agent’s memory file or reset the system, according to the source report. Meta’s help pages describe forgetting as best-effort rather than a guarantee that every trace or inference is removed. Users handling sensitive conversations can avoid connecting those accounts or limit the material they ask Muse to process.
People who do not use Muse have fewer direct controls because they cannot change another person’s agent settings. The report points to WhatsApp’s Advanced Chat Privacy as one way to limit what can be shared from a chat, though it does not prevent all information from reaching an agent through other services or through another participant’s account. Users can also discuss whether messages or documents should be connected to AI tools.
The next significant developments would be clearer documentation from Meta about connector scope, profile creation, retention and deletion, along with any changes to the reported instructions. Meta’s expansion in Singapore makes those details relevant beyond its initial markets. Until the company provides more specific information, users and people in their contacts should distinguish between controls over an agent’s access and a guarantee that information about them will not be processed.
Key Questions
Can Muse read messages from someone who does not use it?
It may process messages available through another user’s connected account. Reports say Muse instructions draw on chats, messages and emails, including information about people mentioned in them. The exact scope depends on the connector and is not fully documented publicly.
Does Muse automatically read every WhatsApp conversation?
No. The source report says Muse in WhatsApp is a chat a user opens, not a reader of that user’s other WhatsApp chats. A conversation could still be accessible through another connected service or account, depending on what its user authorizes.
What can a Muse user do to limit access?
Users can review and disconnect connectors, avoid linking sensitive accounts, change the model-training setting, edit Muse’s memory file or reset the agent. Meta’s materials describe deletion and forgetting as limited; they do not promise that every learned detail will be erased.
Can a non-user delete information from someone else’s Muse?
The source material describes no direct control that lets a non-user manage another person’s Muse. They may be able to limit sharing through settings in specific services, such as WhatsApp Advanced Chat Privacy, or ask the account holder not to connect or share their conversations.
Source: ThorstenMeyerAI.com
