🤖 ⚖🤖 How can we ensure equitable access to AI in Sub-Saharan Africa?
--
Fascinating discussions in 🌴Cotonou, Benin🌴 during the Digital Infrastructure and Cybersecurity conference hosted by Carnegie Mellon University Africa and the CyLab-Africa/Upanzi Network
--with Conrad Tucker, Dr. Eric ADJA and Maximilien F. KPODJEDO, Ph.D
🏡 📩Take-home messages
1) 🌍💬 Securing an African voice in AI development
Training large models from scratch requires resources that are unavailable in resource-limited settings
Open sourcing large models allows open participation and avoids wastefully retraining the wheel 🎡
2) 🛡🛌 Protecting patient safety without stifling AI innovation in Africa
Insist on transparently proven efficacy in the intended population before deployment. [Rigorous real-world clinical studies of AI in healthcare are both safe and innovative💡]
3) 🌐 👑 How to retain data ownership without missing out on the critical benefits of collaboration?
I spoke about one potential solution: building collaborative federated learning to own data ownership
🗽The conference venue was in the shadow of the inspirational "Benin Amazone" statue, honoring the women warriors of the kingdom of Dahomey. The plaque reads: "Like our Amazon warriors of Dahomey, the women of Benin are our pride."