What if your smartwatch could tell when you were struggling emotionally and offer support before you even thought to ask? That's the idea behind a new AI-powered system developed by researchers at the University of Ottawa, one that could change how people access mental health care outside of a clinical setting.
The prototype, called “UbiMyTherapist” (short for "You Be My Therapist"), works as a digital therapy assistant that continuously monitors a user's emotional state and responds with personalized, clinically informed support. It runs on everyday consumer devices like smartwatches, smartphones and earbuds.
The research, led by Dr. Karim Alghoul, part-time professor at uOttawa's School of Electrical Engineering and Computer Science, was conducted under the supervision of a team of two researchers from the Faculty of Engineering: Dr. Hussein Al Osman & Dr. Abdulmotaleb El Saddik, and the help of one student from the School of Psychology, Raina Sharma.
"The core question we asked ourselves was: how can we provide timely mental well-being support outside clinical settings that is safe and personalized to the user?" says Professor Alghoul. "UbiMyTherapist is our answer to that question."
More than a chatbot
What sets UbiMyTherapist apart from standard AI tools is its ability to act proactively. Rather than waiting for a user to type out how they're feeling, the system picks up on emotional cues in real time, through physiological signals like heart rate variability, speech tone, or written text, and responds accordingly.
To do that, it builds and continuously updates what the team calls a "digital twin": a dynamic profile of each user that combines their medical history, a database of clinical psychology knowledge, and live emotional state data. The result is an AI that can tailor its responses not just to what a user says, but to who they are and how they're feeling in the moment.
It draws on three sources of data:
- The user's medical history (such as diagnoses and medications)
- A psychology knowledge database (grounded in clinical literature)
- Real-time emotional state data (detected through biosignals, speech, or text)
By combining all three, the AI generates responses that are not only empathetic and contextually relevant, but also clinically grounded.
The team evaluated a working prototype with volunteers and licensed therapists who were able to assess the system's therapeutic soundness. UbiMyTherapist scored particularly high in empathy and personalization compared to existing LLMs like chatGPT.
"Seeing UbiMyTherapist consistently outperform standard AI setups, especially on empathy and personalization, confirmed that integrating real-time emotional context makes a real difference," explains Alghoul.
“Seeing UbiMyTherapist consistently outperform standard AI setups, especially on empathy and personalization, confirmed that integrating real-time emotional context makes a real difference”
Karim Alghoul
— Part-time professor, uOttawa's School of Electrical Engineering and Computer Science
Filling a gap in mental healthcare
Mental health care remains inaccessible to many people due to cost, stigma, or availability. This research points toward a future where AI can safely bridge that gap, not by replacing therapists, but by extending support beyond clinic walls.
"This work opens the door to accessible, personalized, and context-aware mental well-being support anytime, anywhere," adds Professor Alghoul. "It can also help psychotherapists better understand and support their patients."
The research team plans to expand the prototype to include live proactive interventions triggered by biosignals from the user’s smartwatch in near real-time, and to collaborate further with more licensed psychotherapists to ensure continued clinical alignment.
The study, “UbiMyTherapist: A Digital Twin MultiModal LLM-based System with Emotion Detection”, was published in IEEE (Institute of Electrical and Electronics Engineers). Authorship: Karim Alghoul; Raina Sharma; Hussein Al Osman; Abdulmotaleb El Saddik.