How personal-data access becomes an AI agent’s trust problem
Muse can perform tasks including shopping, email assistance, and trip planning, while drawing personalized information from linked Facebook.
Muse can perform tasks including shopping, email assistance, and trip planning, while drawing personalized information from linked Facebook and Instagram accounts and other interactions.[4] Google’s Dreambeans similarly pulls from Search, Gmail, Photos, Calendar, and Gemini to generate shopping rec…
Muse can perform tasks including shopping, email assistance, and trip planning, while drawing personalized information from linked Facebook and Instagram accounts and other interactions.[4] Google’s Dreambeans similarly pulls from Search, Gmail, Photos, Calendar, and Gemini to generate shopping recommendations, travel suggestions, and reminders.[5]
Why it matters: These agents become more useful by connecting previously separate stores of personal information, but Muse’s ability to surface detailed interests and infer a location from an Amazon shipping address illustrates how that same access can make users reluctant to entrust an agent with email or payment details.[4]
Key insights: Muse successfully handled low-stakes tasks and maintained ongoing lists of suggested uses and goals based on conversations.[4] | Muse said it obtained detailed interest information through Instagram account API data that was more specific than the categories visible in Instagram’s “Ad topics” settings.[4] | The Verge contacted Meta about what Muse can access inside Instagram and Facebook but had not received an immediate response.[4] | Dreambeans is available on Android and iOS after testing began in June.[5]
Cheatsheet facts: What changed: Meta launched Muse, while Google began rolling Dreambeans out to everyone in the US on Android and iOS.[4][5] | Why now: Both products use connected account data to turn general-purpose AI into personalized shopping, travel, communication, and reminder tools.[4][5] | Watch next: Watch for Meta’s explanation of which Facebook and Instagram data Muse can access, following The Verge’s unanswered inquiry.[4]
Muse can perform tasks including shopping, email assistance, and trip planning, while drawing personalized information from linked Facebook and Instagram accounts and other interactions.[4] Google’s Dreambeans similarly pulls from Search, Gmail, Photos, Calendar, and Gemini to generate shopping recommendations, travel suggestions, and reminders.[5]
Why it matters: These agents become more useful by connecting previously separate stores of personal information, but Muse’s ability to surface detailed interests and infer a location from an Amazon shipping address illustrates how that same access can make users reluctant to entrust an agent with email or payment details.[4]
Key insights: Muse successfully handled low-stakes tasks and maintained ongoing lists of suggested uses and goals based on conversations.[4] | Muse said it obtained detailed interest information through Instagram account API data that was more specific than the categories visible in Instagram’s “Ad topics” settings.[4] | The Verge contacted Meta about what Muse can access inside Instagram and Facebook but had not received an immediate response.[4] | Dreambeans is available on Android and iOS after testing began in June.[5]
Cheatsheet facts: What changed: Meta launched Muse, while Google began rolling Dreambeans out to everyone in the US on Android and iOS.[4][5] | Why now: Both products use connected account data to turn general-purpose AI into personalized shopping, travel, communication, and reminder tools.[4][5] | Watch next: Watch for Meta’s explanation of which Facebook and Instagram data Muse can access, following The Verge’s unanswered inquiry.[4]