Digital Phenotyping and Privacy in Mental Health Monitoring

Abstract

Smartphones and wearables now continuously record behavioral data -- movement, communication patterns, sleep, screen use -- that psychiatry increasingly treats as clinically informative, a practice known as digital phenotyping. This review examines how the ethical and legal literature has responded as passive sensing has moved from research protocols into commercial mental health apps and clinical workflows. We survey four connected problems: the mismatch between one-time informed consent and continuous,
purpose-evolving data collection; the gaps in data protection law that leave much of the consumer app market outside meaningful oversight; the error profile and accountability gaps of sensor-derived predictive tools used for relapse or suicide-risk detection; and the psychological effects of living under continuous behavioral observation. Across these themes, the literature converges on a consistent finding: institutional safeguards that partially resolve these concerns in supervised research settings are frequently absent or weaker in commercial deployment. We conclude that consent design, data governance, and algorithmic accountability need to be treated as
core requirements of digital phenotyping tools rather than as compliance steps layered on afterward.