Research Partner - Trustworthy, Human-Centred AI for Healthcare (TOOL-08)

Research Partnership Interest – HORIZON-HLTH-2027-03-TOOL-08

The Accessibility Institute at Carleton University, Canada, is seeking to join a European-led consortium addressing the next generation of frontier AI models for healthcare.

Our research interest lies at the intersection of trustworthy and human-centred AI, cognitive and functional assessment, human-AI collaboration, and AI systems that preserve individual autonomy and agency. We bring interdisciplinary expertise in psychology, accessibility, disability research, artificial intelligence, and machine learning, with particular attention to how AI systems accommodate differences in human abilities, needs, and functional limitations.

Potential contributions to the consortium's coordination, evaluation, and research roadmapping activities include:

• Trustworthy AI and human oversight: Research on multi-agent monitoring, AI safety, transparency, and human-AI collaboration, with particular attention to preserving individual autonomy and decision-making authority. This includes examining how oversight, intervention, and safeguards can remain accessible and effective for people with diverse cognitive, communication, and functional abilities, particularly in high-stakes healthcare contexts.

• Evaluation of advanced AI systems: Developing human-centred evaluation approaches addressing safety, reliability, accessibility, usability, and the effectiveness of AI-supported interactions and decision-making. Particular attention is given to variability in human functioning, disability-related barriers, and the implications of AI errors or limitations for different user populations.

• Inclusive and equitable AI: Examining how frontier AI systems account for human diversity, including cognitive and neurodevelopmental differences, disabilities, and variations in functional abilities. This includes dataset representativeness, bias, equitable access, and the risks of excluding populations whose needs and experiences may be inadequately represented in AI development.

• Person-centred AI applications: Drawing on our research in AI-assisted functional assessment and agency-preserving AI systems to inform future healthcare applications that support accessible communication, understanding of individual needs, and meaningful participation in decisions about care.

• Participatory research and interdisciplinary collaboration: Engaging persons with disabilities, service users, practitioners, and other stakeholders in co-design, evaluation, and the development of research priorities and responsible AI frameworks.

Our research combines technical AI development and evaluation with psychological, social, and ethical perspectives. Our expertise in disability, accessibility, and functional assessment offers a distinctive contribution to the broader challenges of developing trustworthy, inclusive, and human-centred AI for healthcare.

Partnership sought

We welcome discussions with European-led consortia developing a multidisciplinary research and innovation roadmap for frontier AI in healthcare. We are particularly interested in a substantive role involving trustworthy AI, human-centred evaluation, accessible and inclusive AI development, and human oversight, contributing to the call's objectives and the responsible advancement of AI in healthcare.