🔍 Read the full analysis: The Link Between Meta’s AI Agent, Muse, And Your Messages 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 AI agent can access some messages and other data through connectors enabled by its user. Investigations by WIRED and TIME reported that Muse’s instructions tell it to maintain updated pages on people mentioned in users’ communications, including people who never signed up. Meta says each user controls Muse’s access; it remains unclear how nonusers can remove information derived from those conversations.
Meta’s Muse AI agent can access messages and other personal data when its user connects accounts, while instructions reviewed by WIRED and TIME reportedly tell the agent to keep updated pages on people mentioned in those communications. That includes people who have never used Muse, raising questions about their privacy and ability to control information derived from someone else’s chats.
Muse is a personal AI agent that operates on a computer in the cloud and can carry out tasks for a user. Meta says it can connect to services including email, calendars, Mac Messages and Meta’s own apps. The company’s help information says email access lets Muse search messages but does not download the entire inbox unless the user asks. Mac Messages access requires the user to grant macOS Full Disk Access and enable the connector.
Meta’s help page says Facebook, Instagram and Threads connect automatically when accounts 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. Meta’s public help material, however, does not spell out the full scope of what those messaging connectors can read. In WhatsApp, Muse is described as a chat the user opens, not as a reader of their other conversations.
WIRED reported on Muse instruction files obtained by security researcher Karan Joshi, and TIME separately reviewed the instructions. The reports say Muse is directed to create a page for each person in its user’s life and update records based on people named in chats, messages and emails. TIME reported that updates happen hourly and may include relationships, shared interests and disputes. Meta did not dispute the findings in its response to TIME, but said users decide how much access Muse receives.
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 Risks Beyond the User
The issue extends the privacy impact of an AI agent beyond the person who installs it. A user can authorize access to a conversation, but the people in that conversation may not know their words are available to an agent or may be used to build a persistent profile about them.
That distinction matters in family, work and group conversations, where messages can include personal details about several people. Meta’s policy says Muse may store “information about you or others”, while reported instructions describe records about people named in communications. The reports do not establish that every message is always read or that every person receives a profile, but they describe a system designed to retain information about people in the user’s life.
Meta also offers controls to its own Muse users, including disconnecting connectors, turning off model training, editing memory and resetting the agent. Those controls are held by the account owner. People who appear in another user’s messages have fewer direct options, making the practical limits of deletion and opt-out significant.
Products Worth Considering
Logitech MX Master 4 for Mac, Ergonomic Bluetooth Mouse with Haptic Feedback, Ultra-Fast Scrolling, USB-C Charging, for MacOS, iPadOS, Windo
Logitech MX Keys S Wireless Keyboard, Graphite
Apple iPad Air (4th Generation, 10.9-inch, Wi-Fi, 256GB) – Space Gray
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, and introduced it in Japan on October 23, followed by an announcement of Singapore availability on October 27. Meta describes Muse as a personal agent that can answer email, make bookings and shop. It operates in a private virtual machine assigned to each user, which Meta says other users’ agents cannot access.
Muse uses connectors to interact with other services. Some require explicit setup, while Meta says its own Facebook, Instagram and Threads accounts can connect automatically when linked through the same Accounts Center. The company’s privacy policy also describes connectors that can share data proactively. Custom connectors are another possible route; Meta warns that it does not review them.
The report also comes as Meta expands its AI infrastructure. The source material says Meta sold $25 billion in stock at $625 per share and $12.5 billion in preferred stock, with Nvidia investing $3 billion. Nvidia chips are part of Meta’s planned deployment for Muse. Those financing details provide business context, but they do not establish that the investment changes what Muse can access or how its privacy controls work.
“They’re trying to know you like a friend, which is honestly pretty creepy.”
— Karan Joshi, security researcher, to WIRED
Limits of Access and Deletion
The precise scope of message access remains incompletely documented. Meta’s public help pages do not specify exactly what the Instagram and Messenger connectors read, or how often they access data. Meta says email is searchable without downloading an entire inbox unless requested, but connector behavior can differ by service.
The available source material also does not establish how many nonusers have profiles, what information is retained in each case, or how long those records persist. Meta’s help information says Muse may still remember information learned from something a user deleted and describes forgetting as occurring “to the best of its ability.” The source reports that TIME found instructions discouraging the agent from pointing out that original messages may remain visible, but it does not provide a complete account of how deletion works across all connected services.
Meta has not publicly detailed a direct process by which a nonuser can inspect or remove a profile assembled from another person’s conversations. Its response to TIME addressed the control of Muse users, not a separate control for people mentioned in their data.
Controls Users Can Apply Now
Muse users can review connected apps, disconnect connectors they do not need, turn off model training, edit the agent’s memory or reset it, according to the source material. Users should also check permissions for Mac Messages and review any custom connectors, which Meta says it does not review. These steps affect the user’s agent; they do not necessarily remove information already present in another person’s messages or archives.
For people who do not use Muse, the source identifies WhatsApp’s Advanced Chat Privacy as one way to limit what sits in other people’s archives. It does not describe that feature as a general way to erase information from Muse or from other connected services. The next key developments to watch are fuller documentation from Meta about messaging connector access, retention and deletion, and whether the company offers nonusers a way to challenge or remove profiles built from their communications.
Key Questions
Does Muse automatically read every conversation?
The information provided does not say that Muse reads every conversation continuously. Access depends on the connectors and permissions available to the user, and Meta says some tasks draw on connected data when needed. Some connectors may share information proactively, according to Meta’s privacy policy.
Can someone who does not use Muse be included in its records?
Yes, according to WIRED and TIME’s reports on Muse’s instructions. They say the agent is directed to keep pages on people mentioned in its user’s communications, including people who have not signed up.
What can a Muse user do to limit access?
A user can disconnect connectors, turn off model training, edit Muse’s memory or reset the agent. They can also review permissions for services such as Mac Messages and avoid custom connectors they do not trust.
Does deleting a message mean Muse forgets it?
Not necessarily. Meta’s help information says Muse may still remember information learned from deleted material and describes forgetting as happening “to the best of its ability.” The exact deletion process across connected apps is not fully clear.
Can a nonuser ask Meta to remove a profile about them?
The source material does not identify a specific process for nonusers to inspect or remove information Muse may have derived from another person’s messages. Meta’s public response focused on the Muse user’s control over access.
Source: ThorstenMeyerAI.com
