A project led by the University of Barcelona, GenAIMIX, will develop new generative artificial intelligence (AI) assistants for hospitals with limited resources, with the aim of improving healthcare in countries with limited financial resources. The project’s principal investigator is Karim Lekadir, an ICREA researcher at the
UB’s Faculty of Mathematics and Computer Science, head of the
BCN-AIM research group and member of the
Institute of Neurosciences (UBNeuro), which has been awarded a
Proof of Concept grant from the European Research Council. The funding will amount to 150,000 euros.
AI-powered conversational assistants can be a great tool for improving healthcare in resource-poor countries, where a shortage of specialists makes it impossible to carry out high-quality diagnostic imaging. These systems offer intuitive guidance and enable medical and nursing staff with minimal training to acquire images and interpret the results.
However, current generative AI technologies have been developed in high-income countries and are aimed at well-resourced hospitals. This means that these technologies cannot be implemented in low-resource settings due to their computational requirements, their reliance on cloud connectivity, and their inability to operate in local languages.
The GenAIMIX project (Context-Aware Generative AI for Accessible Medical Imaging Across Low-Resource Settings) will develop solutions suited to these resource-limited environments, capable of operating offline on low-power devices. These new technologies will be able to provide real-time guidance to healthcare professionals with limited experience.
The project will be deployed in three rural maternal health clinics in Kenya, where its impact on the acquisition of obstetric ultrasound scans and their use by midwives and nurses will be evaluated. The aim is to expand GenAIMIX’s coverage to other African countries and improve antenatal care through diagnostic imaging.