Decolonized Ethics Framework for AI in Healthcare
AI Ethics
This project aims to develop, test, validate, and valorise a decolonized ethics and governance framework for just and trustworthy AI in healthcare, with regard to the ethical aspects. The Framework will be tailored for Foundation Models (FMs) in the two use cases of freezing of gait (FOG) and rheumatic heart disease (RHD) and DC1 will be in charge of the aspects related to ethics (values and principles). The Framework will be used throughout the project: it will be operationalised through two self-assessment tools respectively focused on FOG and RHD; then, it will be tested and validated through the realization of one pilot action for FOG severity monitoring and one for RHD screening and staging in an every-day life use setting; and it will be eventually valorised, providing insights for standardisation and recommendations for decision- and policy-makers (in collaboration with DC2) as well guidelines for the transferability to other health use cases, providing insights on the ethics-related aspect (in cooperation with DC5, 6, 7&8).
Governance Framework for Just and Trustworthy AI in Healthcare
AI Governance
This project aims to develop, test, validate, and valorise a decolonized ethics and governance framework for just and trustworthy AI in healthcare, with regard to the governance aspects. The Framework will be tailored for Foundation Models (FMs) in the two use cases of freezing of gait (FOG) and rheumatic heart disease (RHD) and DC2 will be in charge of the aspects related to governance (policy and regulation). The Framework will be used throughout the project: it will be operationalised through two self-assessment tools respectively focused on FOG and RHD; then, it will be tested and validated through the realization of one pilot action for FOG severity monitoring and one for RHD screening and staging in an every-day life use setting; and it will be eventually valorised, providing insights for standardisation and recommendations for decision- and policy-makers (in collaboration with DC1) as well guidelines for the transferability to other health use cases, providing insights on the ethics-related aspect (in cooperation with DC5, 6, 7&8).
Co-Creation Strategies for Decolonized AI in Healthcare
Co-Creation Strategies & Methods
This project focuses on how to meaningfully involve patients, healthcare professionals, and other stakeholders in the design of AI-based healthcare solutions. The goal is to develop co-creation strategies and tools that ensure AI systems for specific health conditions are fair, trustworthy, and socially just. DC3 works closely with clinicians, researchers, and communities to identify what is needed for effective collaboration across different contexts. By developing practical co-creation methods and integrating them into the design of AI tools and care pathways, this project ensures that diverse perspectives are embedded from the start. The outcomes support responsible AI development and help translate project results to other healthcare settings worldwide.
Co-Creation Skills for Decolonized AI in Healthcare
Co-Creation Skills
This project focuses on the human skills needed to make AI in healthcare work responsibly and effectively. DC4 investigates which co-creation skills, such as communication, ethical awareness, collaboration, and cultural sensitivity, are essential for patients, professionals, and organisations involved in AI development. The project develops and tests training methods that help stakeholders work together across disciplines and cultures. Through hands-on learning, digital tools, and an international learning community, DC4 supports the development of practical skills needed for fair and trustworthy AI. The results contribute to sustainable capacity building and provide guidance for healthcare organisations and policymakers worldwide.
Deep Learning for Human Movement Modelling
Deep Learning & Human Movement
This PhD project will develop a robust way to assess freezing of gait (FOG) from motion signals that works reliably outside the lab. The core challenge is that models trained in controlled settings often lose accuracy in real-world use, where recordings and daily activities are more variable. The research will address this by improving how models learn general movement patterns and by making them more resilient to changing environments and measurement conditions. It will also focus on turning algorithm outputs into practical, trustworthy insights. Overall, the project bridges advances in AI with real-world usability for patients and clinicians.
Deep Learning for Rheumatic Heart Disease Modelling
Deep Learning & Cardiology
This project aims at developing decolonized Foundation Models for Rheumatic Heart Disease (RHD) screening and staging in resource-limited settings. The research creates a multimodal AI system fusing handheld echocardiography, phonocardiography, and/or electrocardiography data, with built-in bias detection tools, trained on RHD dataset from the Global South. The goal is to enable community health workers, not just expert cardiologists, to perform accurate RHD staging and screening, targeting a high sensitivity and a significant improvement in early detection rates in endemic regions like South Africa.
Just and Trustworthy AI for Freezing of Gait in Parkinson's Disease
Neuroscience & eHealth
This project aims to improve the care of people with Parkinsonian disorders who experience freezing of gait (FOG). We develop a fair, reliable, and scalable care pathway that uses advanced AI technology (foundation models) to better assess and manage FOG in both clinical settings and everyday life. The project focuses on real-world use. We work closely with clinicians, patients, and other stakeholders to co-create solutions that are practical, trustworthy, and socially just. By testing the care pathway in daily life and validating it with clinical experts, we ensure that the technology truly supports people living with FOG.
Just and Trustworthy AI for Rheumatic Heart Disease
Cardiology & eHealth
This project aims to improve the care of those at risk for Rheumatic Heart Disease- a preventable cause of heart disease in low and middle-income countries-affecting over 50 million people and causing 350 deaths annually. DC8 is responsible for developing, testing, validating, and valorising a clinically robust, scalable and just care pathway for Rheumatic Heart Disease (RHD) that integrates Foundation Model (FM)–based screening and staging using handheld echocardiography and phonocardiography (PCG). The project focuses on real-world use. Together with DC6, we will work closely with health care workers, clinicians, patients, and other stakeholders to accelerate diagnosis in community settings.