Apple iPhone depression detection study
The Wall Street Journal reports that Apple, UCLA and Biogen are conducting a study that uses facial recognition, speech patterns and other behavioural data to detect stress, depression and cognitive decline.
First announced in August 2020, the study was initially limited to health data such as heart rate and sleep, and how a person interacts with their iPhone, Apple Watch or Beddit sleep-tracker to understand their mental health.
However, the WSJ's report says the study has been extended to monitor people’s vital signs, movements, speech, sleep, typing habits and frequency of typos, raising concerns from digital rights advocates about the validity of emotion AI/affective computing, privacy, and scope creep.
Apple has recently sought to position itself as a privacy leader, including insisting that apps in the iOS App Store add privacy 'nutrition labels' to inform users what type of sensitive information the app collects.
Developer: Apple; Biogen; UCLA
Purpose: Detect anxiety, depression
Technology: Emotion recognition
Issue: Accuracy/reliability; Pseudoscience; Privacy; Scope creep/nomalisation