University of Ottawa builds proactive AI therapist for wearables

Proactive, sensor-driven mental-health assistants change the data, validation, and privacy requirements for deployed AI. For practitioners, fusing physiological signals with language increases the need for robust multimodal preprocessing, calibration across devices, and stronger privacy-preserving telemetry pipelines. Researchers at the University of Ottawa, led by Dr. Karim Alghoul, built UbiMyTherapist - presented at the 2026 IEEE International Conference on Consumer Electronics - a prototype that uses heart rate variability, speech tone, and written text from smartwatches, smartphones, and earbuds to detect emotional distress and intervene proactively. The system constructs a "digital twin" combining medical and psychological history with live emotional data, operating in both reactive and proactive monitoring modes. Evaluated with volunteers and licensed therapists, it scored higher on empathy and personalization than standard LLM setups. The team plans real-time biosignal response and expanded therapist collaboration; it remains a research prototype, not a consumer product.