12th Dec 2019
How might we...
...inform people about the algorithms used to achieve trust and understanding and give them control
Enabling levels of learning
Smart health devices, products and apps are uniquely placed to enhance people’s lives, but also present difficulty for people to understand how data is collected, stored and used over time. How do digital health services like SleepRoom communicate to people, and allow them to control the algorithms that use their data for personalized health and lifestyle recommendations?
Product Context
In the connected bedroom all devices work together to give you the best possible sleeping experience. SleepRoom automatically adjusts the bedroom’s conditions when it is time to sleep, providing healthy sleep advice and diagnosis on sleep-related conditions.
To provide the service, SleepRoom collects and combines data from 4 sources:
- Smart Watch, to understand a person’s sleeping hours.
- Smartphone sensors, to understand the last time the person uses the phone before going to bed and/or if they use the phone while lying in bed.
- Smart Blanket, to understand body temperature.
- Smart Light, to understand the time window between going to bed and falling to sleep, in addition to the waking up time window.
Problem & Opportunity
SleepRoom currently has a privacy notice that details the app’s data processing activities. The Privacy Notice, however, is very hard to understand for most of their user base. There is an opportunity for SleepRoom to help those using the service, to understand and control how algorithms use their data for personalized health and lifestyle recommendations.
Design Features
Decide how much you want to learn
A key mechanism in the SleepRoom app is that it recognizes people may have different preferences for sharing data. SleepRoom enables people to choose how much they want to share by relating it to the value exchange they will receive from the app. For example, a person can choose the granularity of how much they want to learn and subsequently delve into the data sharing that is needed to support this type of learning.
Design Features
Learn why recommendations are shown
Whenever SleepRoom provides a recommendation it also provides the supporting data points and scientific evidence that backs up the recommendation. Because Sleeproom uses a contained number of data sources it is easier to relate the outputs to something that happened in the data (e.g. recommending to use a smart blanket because body temperature at night is lower than usual).
Next steps
SleepRoom could explore how to refine its recommendations based on the feedback from the people using its products and apps. For instance, when the app recommends a smart blanket to improve sleeping, it can monitor the effectiveness of the intervention and use this information to further tailor and improve its advice.
How might we build on SleepRoom’s ideas to...
- enable a feedback loop between the recommendation and the effect that this has for the user?