Inclusive artificial intelligence using adaptive federated learning across high- and low-resource settings

ongoing

"The AIFIX project is a transformative initiative aiming to bridge the ""healthcare AI divide"" by developing an adaptive federated intelligence platform to securely integrate medical imaging data across diverse environments. It specifically targets the inclusion of low- and middle-income countries (LMICs), where limited data availability, privacy concerns, and inadequate legal frameworks have traditionally hindered AI innovation in healthcare. AIFIX will deliver an adaptive federated learning platform to enable the local training of AI models without the need to transfer raw data, ensuring privacy and compliance with local data protection laws. The platform will be optimised to operate across a variety of computational resource and internet connections, ensuring efficient and stable performance across diverse data centres. The feasibility of this approach will be demonstrated through an inter-continental obstetric imaging study involving diverse sites across Europe, Africa and Asia. From the outset, AIFIX will involve a wide range of stakeholders including healthcare professionals, AI innovators, data managers, and legal experts to ensure the platform meets diverse needs and overcomes potential adoption barriers. Finally, AIFIX will develop a sustainable business model based on detailed market analysis and financial planning. This model will explore various revenue streams such as licensing, subscriptions, and data monetisation, aiming to attract healthcare institutions and AI innovation organisations. The project will also explore federated IP management strategies to protect and share innovations equitably among all contributors. The AIFIX project will represent a significant leap towards equitable AI innovation and utilisation in healthcare, promising to deliver substantial clinical, economic, and societal benefits across the globe."