Background: Despite observational evidence for the efficacy and safety of open-source automated insulin delivery (OS-AID) there is a lack of guidance for healthcare providers. The ethics of providing care for those using unregulated technology is a concern.

Methods: Consensus guidelines are in development, through the OPEN project, by an experienced multi-disciplinary group of endocrinologists, diabetes care and education specialists, exercise physiologists and psychologists.

Results: The principles of biomedical ethics - autonomy, justice, beneficence and non-maleficence - provide a framework for analysis. Autonomy requires that the user understands system risks and benefits. OS-AID provides complete transparency and customization, facilitating understanding and control. Algorithms used in commercial systems are proprietary, impeding the ability to discuss and personalize operation. Justice requires burdens and benefits be distributed fairly. Commercial systems are not uniformly accessible. OS-AID provides freely available software and allows the mixing and matching of compatible CGMs and pumps, supporting equity.

Beneficence means doing good for the individual. In real-world observational trials both open-source (OpenAPS: n=34, mean TIR 71->80% at max 2.5 years; Loop: n=558, mean TIR 67->73% at 6 months) and commercial AID systems (670G: n=3141, mean TIR 66->73% at 3 months; Control-IQ: n=1659, median TIR: 68->78% at 1 month) provide a statistically significant and clinically meaningful increase in TIR at the population level.

Non-maleficence requires treatment does not harm the individual. AID systems can reduce burden, hypoglycemia and hyperglycemia.

Conclusions: The international consensus group concludes OS-AID is ethical, but unregulated. We recommend legal interpretations and frameworks be clarified and updated so that individuals are not deprived of beneficial therapy. Commercial AID systems may benefit from transparency and equity offered by OS-AID systems.