This special collection focuses on enhancing and supporting research related to Sustainable Development Goal 3 (SDG3) and Sustainable Development Goal 10 (SDG10), specifically within the field of healthcare and artificial intelligence (AI).
While the concept of human-centered AI in healthcare has gained traction, concrete evidence regarding its practical outcomes is still lacking. This collection seeks to explore the dimensions of human-AI collaboration in real-world clinical settings, encompassing the perspectives of clinicians, patients, clinical interactions, and broader health systems. By taking a comprehensive sociotechnical approach, we examine how AI influences clinical work, organizations, and healthcare systems, questioning whether it enhances, disrupts, or fundamentally alters these elements. The concept of co-intelligence is a central theme, emphasizing the vision of AI as an intellectual collaborator that aids clinical judgment rather than replaces it. We prioritize original research that is based on empirical data and contributes methodologically, theoretically, or in terms of design, rather than simply presenting descriptive accounts.
Areas of interest include, but are not limited to, the following dimensions, ranging from individual (micro) and care delivery (meso) to societal impacts (macro):
1. The impact of human-AI collaboration on clinician performance, behavior, decision-making processes, and clinical pathways, focusing on how co-intelligence is facilitated through human-in-the-loop workflows where clinicians supervise AI involvement.
2. Human factors that influence clinician-AI interactions, such as automation bias, user-friendliness, and AI’s cognitive impact on clinical decision-making. This also includes examining the long-term effects of AI on clinical expertise and professional development, such as skill acquisition for trainees and skill retention among experienced practitioners.
3. Issues related to professional identity, moral challenges, accountability, and the wellbeing of clinicians working in AI-enhanced environments.
4. Patient experiences with AI in healthcare settings, including interactions with AI-powered decision-making tools and pathways, evaluations of patient-facing AI applications, and the collaborative design of these systems with clinicians and patients. An exploration of how the doctor-patient relationship is evolving in the context of AI is also essential.
5. The design of clinical workflows to optimize AI-assisted diagnosis and care management, clarifying the roles of various clinical professionals in delivering AI-enabled healthcare.
6. Real-world examples of human-AI collaboration, taking into account regulatory, ethical, health equity, economic, and governance perspectives for responsible AI deployment, particularly in diverse and resource-limited environments.
7. Strategies for assessing co-intelligence and its broader impacts on health systems, as well as methods to evaluate and integrate these collaborations into existing healthcare frameworks.
8. Engaging participatory and co-design methodologies throughout the AI lifecycle, enabling the involvement of clinicians, patients, and affected communities in the design, evaluation, and application of clinical AI systems.
We are not seeking research focused solely on model accuracy and will not consider:
- Studies predominantly emphasizing model performance or clinician comparisons. - Simulated case studies where AI-assisted decisions are made in isolation, as these results may not translate to real-world practice. - Usability assessments that fail to provide theoretical, evidential, or design-related insights. - Broad surveys of clinician or patient attitudes towards AI without an exploration of human-AI interaction. - Assertions regarding real-world evaluations of AI's impact on health behaviors that do not systematically analyze the interaction with AI.
We encourage submissions that present multifaceted perspectives, including those from human-computer interaction, information systems, regulatory science, and critical data studies, specifically focused on the foundational aspects of human-AI collaboration. We are particularly interested in research that measures this cooperative dynamic, identifies optimal methods for capturing human-AI interactions, and investigates secondary metrics reflecting the societal implications of AI in clinical contexts.
We welcome a variety of research approaches, including quantitative, qualitative, and mixed-methods studies, with examples such as:
- Quantitative assessments of human-AI collaboration in practice, including stepped-wedge and pragmatic trials, hybrid effectiveness-implementation designs, and interrupted time series evaluations surrounding AI deployment and removal. - Behavior and cognitive studies focusing on real-world contexts, where tasks involve measurable outcomes beyond mere accuracy, including cognitive load assessments and longitudinal analysis of skill growth post-AI withdrawal. - Qualitative work encompassing ethnographic studies, video-based interaction analyses, and user-centered co-design research that includes interviews and iterative qualitative investigations across different stages of AI implementation.
