Consent, Transparency and Data Controls for Alexa Ambient Sensing

Amazon is launching a new ambient sensing feature for Alexa. As the PM, design the privacy controls that ship with it. ```hint Map the data first The right controls depend on what the device detects, where that information is processed and stored, and who in the home is affected. Pin those down before sketching any settings screen. ``` ### Clarifying Questions - What does the feature sense (presence and motion, sounds such as alarms or breaking glass, or daily activity patterns), and with which sensors: microphone, ultrasound, radar or camera? - What happens after a detection: does it trigger routines, send alerts to the owner's phone, or feed personalization? - Can processing stay on the device, or must raw or derived data go to the cloud for the feature to work? - Whose privacy is in scope: the account owner, other household members, children, guests, or neighbors whose sounds might be picked up? ### What a Strong Answer Covers - A clear map of what data the feature collects, where it is processed and stored, and who can see it, used to drive every control. - Controls for each affected person, including household members and guests who never set up the device. - Consent at setup and transparency while sensing is active, with physical and voice controls as well as app settings. - Defaults, retention and deletion choices, weighed against how useful the feature remains. - Metrics that track trust as well as adoption, and risks such as using the feature to monitor another person. ### Follow-up Questions - A household member wants sensing off in their room, but the account owner wants it on. How does your design handle that? - How would you measure whether customers trust the feature, beyond the opt-in rate? - A regulator asks you to show that raw sensor data never leaves the device. What would you need to have built?