Should the Use of Adaptive Machine Learning Systems in Medicine be Classified as Research?

Area of Science:

• Medical Ethics
• Artificial Intelligence in Medicine
• Regulatory Science

Background:

• Adaptive machine learning (ML) systems in medicine can learn from new data post-implementation.
• Current ethical discussions focus on the "update problem" of regulating evolving ML systems.
• A prior ethical consideration regarding continuous learning remains largely unaddressed.

Purpose of the Study:

• To examine whether the continuous learning of medical ML systems post-deployment should be classified and regulated as medical research.
• To argue for a re-categorization of ongoing ML learning in clinical settings.

Main Methods:

• Ethical analysis of adaptive machine learning systems in healthcare.
• Argumentation based on the nature of continuous learning and patient involvement.
• Review of existing regulatory frameworks and ethical principles.

Main Results:

• There is a strong prima facie case that continuous learning in medical ML systems constitutes research.
• Individuals undergoing treatment with such systems should be considered research subjects.
• Current regulatory approaches may be insufficient for adaptive ML systems.

Conclusions:

• Continuous learning in medical ML systems necessitates a re-evaluation of their regulatory status.
• Treating patients as research subjects during adaptive ML deployment ensures ethical oversight.
• This reframing is crucial for responsible innovation in AI-driven healthcare.