The Machine in the Loop – Patient Empowered AI Implementation

Abstract

This case study examines how Type 1 diabetes (T1D) patients and caregivers have leveraged open-source technologies to build more affordable, transparent, and flexible alternatives to commercial insulin management systems. Faced with the high cost, proprietary limitations, and opacity of FDA-approved devices like the MiniMed 670G, a community of patients and developers created tools such as NightScout, OpenAPS, and LoopKit—systems that enable real-time glucose monitoring, algorithmic insulin dosing, and remote care. These innovations offer substantial health benefits, cost savings, and patient autonomy but raise important ethical, legal, and regulatory questions. Developers operate outside FDA frameworks, disclaiming liability and working voluntarily, often driven by personal ties to diabetes. Their systems challenge traditional healthcare delivery by blurring lines between user and developer, patient and engineer. This case invites discussion on transparency, equity, accountability, and legitimacy in biomedical innovation. It asks how regulation, safety, and accessibility can coexist with user-driven innovation, and what role governments, corporations, and open-source communities should play in shaping the future of algorithmic healthcare.