{"meta":{"exportedAt":"2026-10-19T07:41:49.984Z","formatVersion":2,"selection":{"q":"unprompted-message","system":"","harm":"","context":"","country":"","role":"","relation":"core","evidence":"","year":"","response":"","severity":"","verification":"","view":"incidents","sort":"added"},"totalIncidents":1,"coverage":{"cases":1,"countries":0,"languages":1,"unknownLocation":1,"locationPending":0,"unknownLanguage":0,"unknownDate":0,"lawsuits":0,"regulatory":0,"minors":0,"coreRelations":1,"contextualRelations":0,"mixedRelations":0,"unknownRelations":0,"relationPending":0,"relationUnknown":0},"countingNote":"Distinct public cases in this selection. People counts apply within individual cases only; cross-case person overlap has not been resolved. No population incidence estimate.","affectedCountNote":"Interpret person counts with affectedCountStatus and the reported effects. Unquantified zeros are placeholders, not a measured zero.","source":"AI incidents","publisher":"NOPE","url":"https://nope.net/incidents","license":"CC BY 4.0"},"incidents":[{"id":"2026-developer-says-meta-muse-agent-sent-unprompted-message-about-checking-email-meta-blamed-bug-that-caused-hallucination","caseFacts":{"claims":[{"id":"c1","status":"reported","evidence":[{"locator":"'I downloaded Meta's Muse about four weeks ago. I stopped using it three days later.'; 'Three days after downloading Muse, I hopped out of the shower, and I saw a random message pop up from the app.'; 'I can't check your email right now — Gmail isn't connected on my end, so I have no way to look for that Anthropic message.'; 'I hadn't asked for the agent to check my mailbox or messaged it in two days.'","relation":"supports","source_id":"s1"}],"assertion":"The developer says they downloaded Meta's Muse agent about four weeks before the essay of 15 October 2026 and that three days after installing it, and two days after last messaging it, the agent sent an unprompted message saying it could not check the developer's email for an 'Anthropic message' because Gmail was not connected.","causal_attribution":"The developer's account of the agent's message; Meta's later statement (c4) attributes it to a bug that made the model hallucinate."},{"id":"c2","status":"reported","evidence":[{"locator":"'I didn't give it any additional permissions on my device besides permission to access my Facebook, Threads, and Instagram accounts. I never connected it to control anything external, like my email or phone number.'","relation":"supports","source_id":"s1"}],"assertion":"The developer says the agent had been given access only to Facebook, Threads and Instagram and was never connected to email or a phone number.","causal_attribution":"The developer's account."},{"id":"c3","status":"reported","evidence":[{"locator":"'I thought my email might've been hacked.'; 'The whole thing really shook me.'; 'I reported the incident to Meta.'; 'But I disconnected permissions on my Facebook, Instagram, and Threads accounts'; 'I don't use it anymore'","relation":"supports","source_id":"s1"}],"assertion":"The developer says they thought their email might have been hacked and were shaken, reported the incident to Meta, disconnected the agent's permissions and stopped using it.","causal_attribution":"The developer connects the fear and the withdrawal of access to the agent's message."},{"id":"c4","status":"reported","evidence":[{"locator":"'When contacted about this story, a spokesperson for Meta pointed Business Insider to its response on'; 'We investigated with urgency and determined that this was caused by a bug introduced in the product that caused the model to hallucinate and was NOT a security or privacy issue.'; 'The bug has now been fixed.'","relation":"supports","source_id":"s1"}],"assertion":"A Meta spokesperson pointed Business Insider to Meta's reply on the developer's Threads post, which said Meta determined that a bug introduced in the product caused the model to hallucinate, that it was not a security or privacy issue, and that the bug had been fixed.","causal_attribution":"The vendor's own statement as quoted by Business Insider."},{"id":"c5","status":"reported","evidence":[{"locator":"'No damage was done, but what caught me off guard most was that I hadn't done anything to prompt this.'","relation":"supports","source_id":"s1"}],"assertion":"The developer says no damage was done.","causal_attribution":"The developer's own qualification, kept as counterevidence on the extent of harm."},{"id":"c6","status":"documented","evidence":[{"locator":"'Today, Meta is introducing Muse, a secure, private personal AI agent that proactively helps with people's goals and suggests ideas.'; 'Muse is rolling out in the US on iOS, Android, and muse.'; 'September 8, 2026'","relation":"supports","source_id":"s2"}],"assertion":"Meta introduced Muse on 8 September 2026, rolling out in the United States on iOS and Android.","causal_attribution":"Meta's own dated announcement; it bounds the earliest date of the message."}],"effects":[{"label":"fear that an email account had been hacked after an unprompted agent message, leading the user to withdraw the agent's access","claim_id":"c3","direction":"negative"}],"sources":[{"id":"s1","url":"https://www.businessinsider.com/stopped-using-meta-muse-after-strange-message-from-ai-agent-2026-10","kind":"first_person_account","access":"read","language":"en","translation_note":"Read in English on 2026-10-19 from the publisher page; no translation was needed and the research agent (an AI) read the text directly. The essay is the developer's account as told to and edited by Business Insider, with an editor's note quoting Meta's statement.","independence_group":"business-insider-muse-as-told-to"},{"id":"s2","url":"https://about.fb.com/news/2026/09/introducing-muse-personal-ai-agent/","kind":"official_statement","access":"read","language":"en","translation_note":"Read in English on 2026-10-19 from Meta's newsroom; no translation was needed. Used only for the product's introduction date.","independence_group":"meta-newsroom-muse-launch"}],"version":1,"ai_roles":["own_use"],"contexts":["everyday_life"],"unknowns":["The exact date of the message: the developer says Muse was downloaded about four weeks before the essay of 15 October 2026 and the message came three days later, which with Muse's introduction on 8 September 2026 places it in September 2026.","What the bug was and whether other users received similar messages.","The content of the screenshots and of the Threads exchange, which were not inspected.","Whether Meta's investigation reached a further finding beyond the statement quoted."],"geography":{"basis":"Business Insider describes the developer as based in a city in the United States, which places the affected person in the United States; the message arrived in a phone app and no event country is inferred from residence.","court_countries":[],"event_countries":[],"affected_person_countries":["US"]},"publication":{"basis":"Published as a concrete adverse experience with an AI agent: the agent's own unprompted message is quoted, the user describes fearing a hacked email account and withdrawing the agent's access, and the vendor attributes the message to a bug that made the model hallucinate. Claims are attributed to the developer's as-told-to account and to Meta's statement.","reviewed_on":"2026-10-19"},"ai_involvement":{"basis":"The essay quotes the Muse agent's own unprompted message saying it could not check the user's email for an 'Anthropic message' because Gmail was not connected, and Meta's statement says a bug introduced in the product caused the model to hallucinate. The user connects that message to fearing the email account had been hacked, reporting the incident and disconnecting the agent's permissions.","status":"supported"},"person_relations":["communicated_with"]},"name":"Developer says Meta's Muse agent sent an unprompted message about checking an email account it could not access; Meta said a bug caused a hallucination","summary":"In an as-told-to essay published by Business Insider on 15 October 2026, a software developer says that three days after installing Meta's Muse agent, and two days after last using it, the agent sent an unprompted message saying it could not check the developer's email for an 'Anthropic message' because Gmail was not connected. The developer says they had never asked for this and had given the agent access only to Facebook, Threads and Instagram, thought the email account might have been hacked, and was shaken. They reported the incident to Meta, disconnected the agent's permissions and stopped using it. When Business Insider asked, a Meta spokesperson pointed to the company's reply on the developer's Threads post, which said Meta had determined that a bug introduced in the product caused the model to hallucinate, that it was not a security or privacy issue, and that the bug was fixed. The developer says no damage was done. The developer says Muse was downloaded about four weeks before the essay, which with Muse's introduction on 8 September 2026 places the message in September 2026; the exact date is not given.","incidentDate":"2026-09-01","incidentKind":"single_event","incidentDatePrecision":"month","exposurePattern":"single_interaction","reportedDate":"2026-10-15","aiSystem":"Meta Muse personal AI agent (phone app) connected to the user's Facebook, Threads and Instagram accounts","aiProduct":"Meta Muse agent","aiCompany":"Meta","severity":"low","verificationStatus":"credible","harmCategories":[],"harmOutcomes":["psychological_distress"],"harmOutcomeSummary":"In a Business Insider as-told-to essay, a software developer says an unprompted message from Meta's Muse agent about checking an email account it had no access to made them fear the account had been hacked and shook them; they reported it, disconnected the agent's permissions and stopped using it. Meta said a product bug made the model hallucinate.","frameworkFacets":[],"causationStatus":"supported","participantUsersAffectedMin":1,"otherPeopleHarmedMin":0,"affectedCountStatus":"exact","affectedCountEvidence":"One person: the developer who received the agent's message and tells the story in the essay. Exact 1.","victimAgeRange":"adult","jurisdiction":"unknown","platformType":"agent","outcomeType":"internal_action","outcomeStatus":"resolved","primarySourceUrl":"https://www.businessinsider.com/stopped-using-meta-muse-after-strange-message-from-ai-agent-2026-10","primarySourceLabel":"Business Insider, 15 October 2026: I tried Meta's Muse AI agent for 3 days, then it sent me a strange message (as-told-to essay)","firstPublishedAt":"2026-10-19T03:13:17.436223+00:00","updatedAt":"2026-10-19T03:13:17.436223+00:00","scopeVersion":"facts-v3","tags":["ai-agent","unprompted-message","hallucination","vendor-acknowledged","as-told-to"]}]}