🔍 Read the full analysis: Does Meta’s AI Agent Read Your Messages Without Muse? on ThorstenMeyerAI.com
Get business pricing on tech for your team
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
Meta’s Muse personal AI agent can use connected services, including some Meta apps, to search and act on information available to its user. Reports by WIRED and TIME say Muse’s instructions tell it to maintain frequently updated pages about people in that user’s life, including people who never signed up. Meta says users control the agent’s access; how non-users can remove information inferred from other people’s messages remains unclear.
Meta’s Muse personal AI agent can search information its user makes available through connected services, and reports by WIRED and TIME say its instructions direct it to keep updated pages about people named in that user’s messages and emails. That may include people who never signed up for Muse; Meta says each user controls how much access the agent receives, but the reports raise questions about what control those other people have.
Muse uses connectors to link to other services. Meta’s help information says an email connector lets Muse search email, but does not download the entire inbox unless the user tells it to. Calendar connectors may send updates proactively. Meta also says Facebook, Instagram and Threads connect automatically when accounts share an Accounts Center with Muse. The source material says the scope of those connections is not fully documented in Meta’s help pages.
Meta chief AI officer Alexandr Wang said in a launch-day post that the Instagram connector includes direct messages and that Messenger is also supported. Muse can connect to Apple’s Mac Messages app if the user grants Full Disk Access and enables the connector. Meta says Muse can also be used in WhatsApp as a chat a person opens; the source material distinguishes that interaction from Muse reading the user’s other WhatsApp chats.
TIME and WIRED reported that Muse’s internal instructions call for “a page for every person in the user’s life” and for regular updates to information about people mentioned in chats, messages and emails, as detailed in reports about Muse reading messages. The source account says TIME reviewed the instructions separately after WIRED’s report, and that Meta did not dispute TIME’s findings. Meta’s spokesperson said: “Each person stays in control of their Muse and decides how much access it gets,” though Muse’s trust concerns remain.
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
Privacy Beyond the Muse User
The reports point to a privacy gap between the person using an agent and the people they communicate with. A Muse user may authorize access to a mailbox or social account containing messages from friends, relatives or colleagues. If the agent then extracts details about those people into a profile, those details may be processed even though the people described never created a Muse account.
That matters because an AI agent is designed to retrieve and act on information, not simply display it. A profile that collects relationship history, shared interests or unresolved disagreements could shape later summaries and responses. The source material reports that Muse’s instructions call for updates and describe shared, de-identified “lessons” across the system. The precise operation and effects of that sharing are not established in the material provided.
Control also appears uneven. Muse users can change their own access settings, while non-users may have fewer direct ways to affect information derived from someone else’s archive. Meta’s statement addresses the user’s choices, but does not explain a separate process for people mentioned in those communications.
Products Worth Considering
Logitech MX Brio Ultra HD 4K Webcam, Graphite
DJI Mic 3 Wireless Microphone System (2 Transmitters, 1 Receiver, Charging Case)
Apple iPad Air (2025) 13‑inch Wi‑Fi 256 GB – Space Gray (M3 chip)
As an affiliate, we earn on qualifying purchases.
Muse’s Connectors and Rollout
Meta introduced Muse in the United States on September 8, describing it as a personal AI agent that can work through a cloud-based computer called a Muse Secure VM. The company says one user’s agent cannot reach another user’s VM. Japan followed on October 23, and Meta announced a Singapore launch on October 27, according to the source material. TIME reportedly counted more than five million downloads within weeks; the material does not provide a measurement method or a comparison baseline for that figure.
The concerns are tied to the agent’s access model. Meta’s own policy, as summarized in the source material, says Muse may store “information about you or others.” Users can disconnect connectors, turn off model training, edit a memory file or reset the agent. The material says model training is on by default, but does not provide further detail about the setting’s precise scope. Meta also warns that it does not review custom connectors, which users can create to link Muse with other services.
The source account also describes a large financing and infrastructure development: Meta sold $25 billion in stock at $625 a share and $12.5 billion in preferred stock, while Nvidia said it would invest $3 billion. Meta plans to use Nvidia chips as the companies expand their work on Muse. Those figures provide business context, but do not by themselves establish how the investment will affect Muse’s privacy practices.
“They’re trying to know you like a friend, which is honestly pretty creepy.”
— Karan Joshi, security researcher, quoted by WIRED
What Non-Users Can Remove
The available material does not establish exactly what information each connector collects, how long derived profiles are retained, or whether a non-user can view or delete them. It also does not explain how frequently information from every connector is read in practice, beyond the reported instruction for hourly updates to certain files.
Meta’s help guidance, as quoted in the source material, says Muse may still remember information it learned after a user deletes something, and that it forgets information “to the best of its ability.” That leaves open what deletion means for details inferred from messages, and whether removing a source message also removes a derived profile. The source account says TIME reported that Muse’s instructions do not prompt it to tell users that original messages may remain visible; Meta’s public explanation of that point is not provided here.
The source material says WhatsApp’s Advanced Chat Privacy can limit what sits in other people’s archives, but it does not establish how effective that setting is across all apps or whether it prevents every form of processing. The reports’ descriptions of Muse’s internal instructions are attributed findings, not a full public technical audit.
Meta’s Privacy Details to Watch
The next useful information would be a clear, connector-by-connector account from Meta of what Muse reads, what it stores, how long it keeps profiles about third parties and how users can remove both source data and information inferred from it. A separate explanation of what options are available to people who never signed up would address the central gap raised by the reports.
Users can review which connectors are enabled, disconnect services they do not want Muse to access, check model-training settings and review or reset the agent’s memory, according to the source material. People who do not use Muse may be able to limit some information shared through individual services, but the material does not describe a universal opt-out. Meta’s response to the specific questions about non-users, retention and deletion remains unclear.
Key Questions
Does Muse read every message in a connected inbox?
Meta’s help information says an email connector lets Muse search email but does not download the whole inbox unless the user tells it to. The exact scope and timing can vary by connector, and the source material says some connector data may be shared proactively.
Can Muse access Instagram DMs and Messenger?
Meta chief AI officer Alexandr Wang said Muse has an Instagram connector that includes direct messages and a Messenger connector. The source material says Meta’s help pages do not spell out the full scope of those connections.
Can someone who has never used Muse be included in its records?
WIRED and TIME reported that Muse’s instructions tell it to keep pages about people in its user’s life and update information about people mentioned in communications. That means non-users may be described in an agent’s records, according to those reports.
How can a Muse user limit the agent’s access?
The source material lists disconnecting connectors, turning off model training, editing the memory file and resetting Muse. Users can also review the permissions they granted to services such as Mac Messages.
Can non-users delete information Muse has inferred about them?
The material does not establish a direct deletion or opt-out process for non-users. Meta’s statement says the Muse user controls the agent, while questions about removing information derived from other people’s messages remain unanswered.
Source: ThorstenMeyerAI.com
