Artificial intelligence in clinical decision-making: Rethinking personal moral responsibility

Area of Science:

• Medical Ethics
• Artificial Intelligence in Healthcare
• Software Development

Background:

• Clinicians traditionally lead healthcare decisions.
• Software developing companies (SDCs) are introducing artificial intelligence systems (AISs) to influence clinical decision-making.
• SDCs state clinicians remain the final decision-makers, with AISs only informing choices.

Purpose of the Study:

• To examine the attribution of personal moral responsibility for outcomes involving AIS in clinical decision-making.
• To challenge the default position that clinicians solely hold responsibility.
• To propose a framework for just responsibility attribution.

Main Methods:

• Analysis of prospective versus retrospective responsibility.
• Examination of foreseeability as a key factor in attributing moral responsibility.
• Ethical examination of SDC and clinician roles in AIS use.

Main Results:

• The current default position of clinician-only responsibility is imbalanced and unjust.
• Foreseeability is crucial for justly attributing moral responsibility.
• AISs influence clinician judgment, complicating sole clinician accountability.

Conclusions:

• Moral responsibility for AIS outcomes in healthcare should be shared between clinical users and SDCs.
• A nuanced approach considering foreseeability is necessary for ethical AIS implementation.
• This shared responsibility model promotes a more just and balanced approach to AI in clinical settings.