The problem
Organizations that care for people with developmental disabilities and older adults face a growing shortage of direct care staff and high staff turnover, while the number of people who need support keeps rising. Too much scarce staff time goes to routine checks and monitoring instead of to the people in their care. We asked whether a robot could safely and respectfully take on some of that routine work, so staff could spend more time with residents.
What we built
Over ten years, JDQ led the research using its 3Spheres™ model. It brought together non-profit care providers, university researchers from UBC (including the UBC IDEA Lab), the University of Toronto and Ontario Tech University (OTU), and funders including Mitacs, NSERC and AGE-WELL, all working under a project code of ethics. Early work included focus groups in 2018 that compared the needs of care staff in group homes and seniors' care homes, and found many similarities. The team then designed and prototyped Aether™, a collaborative service robot with autonomous navigation and mapping, person and face recognition, pose and fall detection, and voice interaction. Its skills included safety inspections that report hazards, going to a named location, and entertainment that helps calm and redirect residents. When COVID-19 closed care homes to visitors, much of the design and testing moved into simulation and a virtual-reality model of a long-term care facility. Aether™ was field tested in a group home and in a 100-bed residential care facility, and was shown publicly in Vancouver in 2024. The project was put on hold indefinitely in June 2025.
JDQ won the 2007 BCTIA Technology Impact Award for Leadership in Social Responsibility.Results
Lessons
3SRP gave JDQ a deep, practical education in AI, machine learning and large language models, applied where the stakes are highest: health care, care for older adults and care for people with developmental disabilities. It reinforced why AI needs a graduated trust model, where a system earns more independence only as it proves itself safe and reliable, and people stay in charge. Aether™'s deployment stalled when AI began advancing faster than care organizations and their users could ensure the safety and privacy of the people using it. AI holds great promise for these sectors, but for organizations not yet equipped to manage its risks, the potential for harm is simply too high. Those lessons guide how JDQ applies AI today, and two of our tools for reducing AI risk grow directly out of them. Our AI Readiness Assessment checks that an organization can manage AI's risks to safety and privacy before anything is deployed. And the graduated trust model is built into every agent-first app we deliver, so each agent's independence matches the trust it has earned, and people stay in charge.