ExplainabilityTransparencyTrustworthinesseess.SP
Visualisation to Explain Personal Health Trends in Smart Homes
Glenn Forbes, Stewart Massie, Susan Craw
September 28, 2021
1 citations
Key Takeaways
-
Transparency builds trust in health monitoring systems: The paper demonstrates that when AI systems make health recommendations based on sensor data in smart homes, residents are more likely to trust these systems when the decision-making process is made visible and understandable, rather than presenting recommendations as "black boxes."
-
Visual explanations can make complex health data accessible: The proposed color-coded graph visualizations translate raw sensor data and activity profiles into formats that non-technical users can comprehend, addressing a key barrier to responsible AI deployment in healthcare settings.
-
Sensor data requires careful interpretation: The research highlights that ambient sensors in smart homes generate activity profiles that are then compared against residents' personal baselines and "risky" profiles to recommend interventions—a process that benefits from explainability because the stakes (health recommendations) are significant.
-
Explainability is essential for health AI in intimate settings: Because smart home monitoring occurs in private residential spaces and directly influences health decisions, the paper underscores that XAI (explainable AI) isn't optional but necessary for maintaining resident autonomy and informed consent.
Abstract
An ambient sensor network is installed in Smart Homes to identify low-level events taking place by residents, which are then analysed to generate a profile of activities of daily living. These profiles are compared to both the resident's typical profile and to known "risky" profiles to support recommendation of evidence-based interventions. Maintaining trust presents an XAI challenge because the recommendations are not easily interpretable. Trust in the system can be improved by making the decision-making process more transparent. We propose a visualisation workflow which presents the data in clear, colour-coded graphs.