🔍 Read the full analysis: What Meta’s AI Agent May See Even If You Never Open Muse 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 access information through connected services, and reports by WIRED and TIME say its instructions direct it to maintain profiles of people mentioned in users’ messages. Meta did not dispute TIME’s account, but says each Muse user controls the agent’s access. The scope of data retention and what non-users can do to limit profiling remain unclear.
Meta’s Muse personal AI agent may compile information about people who never signed up, based on reports that its internal instructions tell it to maintain pages about people mentioned in a user’s chats, messages and emails. WIRED and TIME reported on those instructions in October; Meta did not dispute TIME’s findings and said each user controls what Muse can access, leaving questions about the privacy choices available to people whose information appears in someone else’s accounts.
Muse accesses services through connectors, which let the agent work with information in linked accounts. Meta’s help pages say users can connect email and calendars, while Facebook, Instagram and Threads connect automatically when those accounts share an Accounts Center with Muse. Meta chief AI officer Alexandr Wang said the Instagram connector includes direct messages and that Messenger is also supported. Meta’s help pages, as described in the source report, do not fully document what those connectors read.
For email, Meta says a connection lets Muse search messages but does not download the entire inbox unless the user asks. Calendar connectors can send updates proactively. Access to Apple’s Mac Messages requires the user to grant macOS Full Disk Access and enable that connector. Meta also warns that it does not review custom connectors, which users can create for other services. These controls apply to the Muse user’s account; they do not provide a direct setting for another person whose messages appear there.
TIME reported that Muse’s instructions call for hourly updates to files about the user and people named in the messages it can read. The reported profiles may include relationships, shared interests, disputes and group dynamics. WIRED had earlier reported an instruction to create a page for every person in the user’s life. The reports describe internal instructions, not an independently verified account of how often every profile is created or what information is retained in practice. Meta’s policy says Muse can store information about the user or others.
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 issue extends an AI privacy decision beyond the person who installs the product. If a Muse user connects accounts containing conversations with family, colleagues or friends, those contacts’ information may become available to the agent as part of the user’s activity. They may not know that the account is connected, and they cannot use the same connector settings to control what the agent sees.
That matters especially for group chats and private messages, which can contain details about multiple people, not just the account holder. Meta says users decide how much access Muse receives, but that assurance describes the user’s control rather than the choices available to everyone mentioned in the data. The distinction affects how people assess consent, retention and the practical meaning of deleting or asking an agent to forget something.
The commercial stakes are also relevant but do not establish a motive for collecting any particular information. The source reports a $25 billion stock sale at $625 a share, alongside $12.5 billion in preferred stock, and says Nvidia plans a $3 billion investment as Meta expands work on Muse. Meta is also deploying Nvidia chips for the agent, according to the source. Those financial developments show the scale of investment around Muse, but do not by themselves prove how data will be used or how the product will make money.
Products Worth Considering

Apple iPad Air (2025) 13‑inch Wi‑Fi 256 GB – Space Gray (M3 chip)

Logitech MX Master 4 for Mac, Ergonomic Bluetooth Mouse with Haptic Feedback, Ultra-Fast Scrolling, USB-C Charging, for MacOS, iPadOS, Windo

Sonos Era 100 SL Compact Microphone-Free Speaker, Black
As an affiliate, we earn on qualifying purchases.
Muse’s Launch and Data Connections
Meta launched Muse in the United States on September 8, 2026, then expanded to Japan on October 23 and announced Singapore on October 27. Meta describes it as a personal agent that can act on a user’s behalf, including handling tasks such as email and bookings. TIME reported more than five million downloads within weeks; that figure is TIME’s reported estimate, not a number independently detailed in the source material.
Meta says each customer receives a Muse Secure VM, a virtual computer that the company says is inaccessible to other users’ agents. That is a claim about separation between agents; it does not answer what information a user’s own agent can access through enabled connectors or how long information about other people may remain in its records. Muse can also be opened as a chat in WhatsApp, which the source says is separate from reading the user’s other WhatsApp chats.
The reported profiling instructions surfaced through two investigations. On October 3, WIRED reported on files that security researcher Karan Joshi said he obtained by asking Muse in its chat. TIME reviewed instructions separately three days later. The source report says Meta did not dispute TIME’s findings. Meta’s help materials also say users can disconnect connectors, turn off model training, edit a memory file or reset Muse, although the exact effect of each step on information already derived about other people is not fully established.
“They’re trying to know you like a friend, which is honestly pretty creepy.”
— Karan Joshi, security researcher, quoted by WIRED
What The Reports Do Not Establish
The reporting describes Muse’s instructions, but the available material does not establish how consistently the agent creates or refreshes profiles, whether every connected message is used in that process, or how long those records persist. Meta did not dispute TIME’s findings, but its public response, as provided here, does not answer those operational questions or explain what information may be shared as de-identified lessons across the system.
It is also unclear how much a user’s controls change information already inferred about other people. Meta’s help language says Muse forgets information “to the best of its ability” and may retain learned information after a user deletes the source. The source report says TIME found instructions telling Muse not to point out that original messages may remain visible. The full policy wording and the effect of deletion, reset and training settings on third-party information are not supplied in full here.
The financial figures and product rollout are reported in the source material, but they do not show that Nvidia’s investment is tied to particular data practices. Nor does the availability of connectors prove that Muse reads every connected account continuously: Meta’s descriptions distinguish task-based access from certain proactive updates. The precise scope and timing of access remain insufficiently documented in the material.
Controls Users Can Change Now
Muse users can review connected services, disconnect connectors they do not want the agent to use, and check whether Facebook, Instagram and Threads are linked through the same Accounts Center. They can also review the model-training setting, edit Muse’s memory file or reset the agent. Those steps can reduce the information available to a user’s agent, though they may not remove copies or inferences already retained elsewhere.
People who do not use Muse have fewer direct controls. The source points to WhatsApp’s Advanced Chat Privacy as one way to limit what other people can export or use from a chat, but it does not establish that this blocks every form of access through another person’s connected account. Users should check the current controls and app documentation rather than assume a privacy setting applies across all services.
The next important developments are clearer documentation from Meta about profile creation, retention, deletion and third-party controls, along with scrutiny of the product as it expands beyond North America. Meta announced Singapore on October 27; the source does not provide a date for wider availability or a detailed response to the specific questions raised by the reports.
Key Questions
Can Muse access messages from someone who never signed up?
Potentially, through another user’s connected accounts. Reports say Muse’s instructions cover people named in messages it can read. A non-user does not directly grant Muse access, but their messages may be present in an account connected by someone else.
Does Muse automatically connect to Instagram and Facebook?
Meta’s help information says Facebook, Instagram and Threads connect automatically when they share an Accounts Center with Muse. Meta chief AI officer Alexandr Wang said the Instagram connector includes DMs and that Messenger has a connector, though the supplied help details do not fully define what those connections read.
Does asking Muse to forget something delete it?
Not necessarily. Meta’s help language, as quoted in the source, says Muse forgets information “to the best of its ability” and may still remember information learned from deleted material. The exact effects of deletion or reset on information about other people are unclear.
What can a person who does not use Muse do?
They have fewer direct options than a Muse user. They can review privacy features in the apps they use; the source names WhatsApp’s Advanced Chat Privacy as one possible limit on how chat content is handled. The material does not show that this prevents all access through another person’s account.
Has Meta confirmed that every user gets a profile of every contact?
No. WIRED and TIME reported instructions directing Muse to create and update pages about people in a user’s life, and Meta did not dispute TIME’s findings. The reporting provided here does not establish how consistently those instructions are carried out or whether every person mentioned receives a profile.
Source: ThorstenMeyerAI.com
