AIMIX · Inclusive Artificial Intelligence for Accessible Medical Imaging Across Resource-Limited Settings

Artificial intelligence (AI) is widely regarded as one of the most promising and disruptive technologies for future healthcare. As AI algorithms such as deep neural networks are suited for the processing of large and complex datasets, radiology is the medical speciality that has seen some of the most important applications of AI in the recent years. However, despite these advances, a major limitation of current AI developments in medical imaging is that they have overwhelmingly, and almost entirely, targeted applications in high-income countries. There is a concern, if the current trend continues, that AI will increase the already pronounced inequalities in global health, in particular for resource-limited settings such as rural Africa, where the majority of the African population lives.AIMIX will develop the first scientific framework for inclusive imaging AI in resource-limited settings. The project will greatly advance the current state-of-the-art, from existing AI methods mostly developed for high-income settings, towards new imaging AI algorithms that are fundamentally inclusive, i.e. (1) affordable for resource-limited clinical centres, (2) scalable to under-represented population groups, and (3) accessible to minimally trained clinical workers. Furthermore, AIMIX will investigate the socio-ethical principles and requirements that govern inclusive AI, and examine how they compare, conflict or complement those of trustworthy AI developed thus far in high-income settings. These innovations will be demonstrated for affordable and accessible AI-powered obstetric ultrasound screening by minimally trained clinicians such as midwives in rural Africa.Ultimately, AIMIX’s scientific breakthroughs will enhance the democratisation of imaging AI in resource-limited settings, which will result in an important social impact, by empowering local communities, promoting inclusion, and reducing disparities between populations from low- and high-income societies.