🔍 Read the full analysis: Meta’s AI Agent May Read Messages: What Muse Has To Do With It on ThorstenMeyerAI.com
Get smart everyday buys 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 expanded its Muse personal AI agent to Singapore on Oct. 27, while reports described instructions that direct the agent to maintain profiles of people named in a user’s messages. Meta says Muse users control its access, but people who never signed up may still be described in information their contacts share. The precise scope of those profiles and how deletion works remain unclear.
Meta announced on Oct. 27 that its Muse personal AI agent is coming to Singapore, as reports about the agent’s handling of connected messages raise questions for people who have never used it. Meta says each user controls what Muse can access; WIRED and TIME reported that its instructions direct it to keep and update pages about people mentioned in users’ communications.
Muse is a cloud-based assistant that can perform tasks such as answering email, booking tables and shopping. Meta says it runs in a separate virtual computer for each user. The company says people can interact with Muse through its app, on a Mac or in a dedicated WhatsApp chat. That WhatsApp chat is not described as giving Muse access to a user’s other WhatsApp conversations.
Access to other services depends on connectors. Meta’s help information says email access lets Muse search email, but does not download an entire inbox unless the user asks. A calendar connection can send updates to Muse. Facebook, Instagram and Threads accounts may connect automatically when they share an Accounts Center with Muse. Meta chief AI officer Alexandr Wang said the Instagram connector includes direct messages and that Muse has a Messenger connector; the source material says Meta’s help pages do not document the full scope of what those connectors read.
Mac Messages access requires the user to grant macOS Full Disk Access and enable the connector, according to Meta. Users may also create custom connectors, but Meta warns it does not review them. Meta’s privacy policy describes some connector data as shared proactively and lists messages, connector data and information about other people among the information it may collect.
WIRED reported on instructions obtained by security researcher Karan Joshi that tell Muse to create “a page for every person in the user’s life.” TIME later reviewed instructions that reportedly direct hourly updates to files about the user and people mentioned in chats, messages and emails. Those files may include details about relationships, shared interests and disagreements, according to TIME. Meta did not dispute TIME’s findings in the account provided here.
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 Become Data
The issue extends beyond the person who installs Muse. If an agent can access a user’s messages, details about the people in those conversations may be available for the agent to summarize or retain. That could include people who did not choose Meta’s AI service and may not know their messages are being processed by a contact’s assistant.
The distinction is between control over an account and control over information about oneself. Meta says the Muse user decides what access to grant. People whose information appears in that user’s communications do not have the same direct settings in the other person’s account. The reported profile-building instructions make that gap a practical privacy concern, although the source material does not establish how often such profiles are created or what information they contain in real-world use.
The expansion also matters because Muse’s usefulness depends on access to personal services and communications. Meta is developing the agent alongside a major financing and infrastructure effort: the company sold $25 billion in stock at $625 a share and $12.5 billion in preferred stock, while Nvidia said it would invest $3 billion. The source describes Nvidia as a supplier of chips Meta will deploy for Muse. Those figures show the commercial scale around the rollout, but do not by themselves establish how Muse handles any individual’s data.
Products Worth Considering
Dell XPS 13 9350 AI Professional Laptop 13.4 120Hz FHD+ Display
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
As an affiliate, we earn on qualifying purchases.
Muse’s Rollout and Data Access
Meta launched Muse in the United States on Sept. 8, 2026, and expanded it to Japan on Oct. 23, before announcing Singapore on Oct. 27. Meta described the product as “the world’s first personal AI agent built for everyone.” That is the company’s characterization, not an independently established comparison. TIME reported more than five million downloads within weeks of launch; the source material does not provide a comparison baseline or further detail on how that figure was measured.
Muse’s connectors are the route through which it works with outside services. Some require an explicit setup, while Meta says Facebook, Instagram and Threads can connect automatically when accounts share an Accounts Center. Meta’s own materials also describe data that may be shared proactively. The details vary by connector, and the source material says the public help pages do not explain the full scope of Instagram DM and Messenger access.
Meta’s help information says users can disconnect connectors, switch off model training, edit a memory file or reset Muse. It also cautions that after deletion, Muse may still remember information it learned, “to the best of its ability.” This wording means that a request to forget information should not be treated as a guarantee that all traces or derived memories are removed.
“They’re trying to know you like a friend, which is honestly pretty creepy.”
— Karan Joshi, security researcher, quoted by WIRED
Open Questions on Profiles and Deletion
The available information does not establish how many people have been profiled by Muse, how detailed any individual profile becomes, or how long information about non-users is retained. Meta did not dispute TIME’s account of the instructions, but the source material does not include a full technical explanation from the company of how those instructions operate in practice.
It is also unclear exactly which message content each connector can access, when it reads that content, and whether a user’s settings affect every kind of information derived from it. Meta says users control access, while its help information acknowledges that Muse may retain learned information after deletion. The source does not explain what data persists, for how long, or whether a non-user can request removal of information held in someone else’s Muse account.
What to Watch as Muse Expands
Meta’s announced Singapore rollout is the next market development. The company has not, in the source material, supplied a date for broader availability there or a fuller public explanation of how its reported person-profile instructions work. More detail about connector permissions, retention and deletion would help users and people in their communications understand the practical limits of the controls.
For current users, the available steps described by Meta include reviewing and disconnecting connectors, switching off model training, editing Muse’s memory file or resetting the agent. People who do not use Muse have fewer direct controls over information in another person’s account. The source material points to WhatsApp’s Advanced Chat Privacy as one way to limit what sits in other people’s archives, but does not establish that it prevents all copying or processing across every service.
Key Questions
Can Muse read Instagram direct messages?
Meta chief AI officer Alexandr Wang said the Instagram connector includes direct messages. The source material says Meta’s public help pages do not spell out the full scope of what the connector can access.
Does Muse automatically read every WhatsApp conversation?
No such access is described in the source material. Meta says Muse can be opened as a dedicated WhatsApp chat; that is distinct from access to a user’s other WhatsApp chats.
Can Muse create information about people who do not use it?
WIRED and TIME reported that Muse instructions direct it to maintain pages about people mentioned in a user’s communications. That means non-users may be described in information handled through another person’s Muse account, but the source does not establish how often this occurs or the precise data retained.
Does deleting a memory guarantee that Muse forgets it?
Meta’s help information says Muse may still remember information it learned after deletion and describes forgetting as happening “to the best of its ability.” The available information does not specify what may remain or for how long.
What can a Muse user do to limit access?
Meta lists disconnecting connectors, turning off model training, editing the memory file and resetting Muse as available controls. Users should review connector permissions, particularly where an account is connected automatically or a custom connector is involved.
Source: ThorstenMeyerAI.com
