A theoretical approach to artificial intelligence systems in medicine

Area of Science:

• Medical Informatics
• Artificial Intelligence in Medicine
• Philosophy of Medicine

Background:

• The development and application of Artificial Intelligence (AI) systems in medicine are intrinsically linked to prevailing theoretical models of disease, diagnostic logic, and established nosology.
• Implicit theoretical assumptions and socio-cultural attitudes embedded within AI systems can significantly influence diagnostic and therapeutic outcomes.
• Existing AI approaches in healthcare often lack a robust, self-consistent theoretical framework, reflecting the complexity of human biology and the non-explicit nature of AI's underlying assumptions.

Purpose of the Study:

• To critically examine various theoretical models underpinning AI in medicine, including causal, probabilistic, and case-based approaches.
• To elucidate the ethical and methodological limitations inherent in current AI systems used for medical diagnosis and treatment.
• To explore the implications of these limitations on the role of AI in clinical decision-making and medical education.

Main Methods:

• Critical analysis of theoretical models of disease and diagnosis.
• Examination of the implicit assumptions and socio-cultural factors influencing AI in medicine.
• Evaluation of the ethical and methodological constraints of AI systems in healthcare.

Main Results:

• Theoretical models and diagnostic logic shape medical approaches and AI development.
• AI systems in medicine incorporate implicit assumptions that bias outcomes.
• Various AI models (causal, probabilistic, case-based) exhibit ethical and methodological limitations.
• The complexity of medicine and implicit AI assumptions restrict AI to a decision-support role, not autonomous decision-making.

Conclusions:

• AI systems currently function as decision-support instruments rather than autonomous decision-making devices due to theoretical gaps and inherent complexities.
• The crucial role of AI in medical education necessitates further investigation into its theoretical underpinnings and developmental trajectory.
• Advancing AI in medicine requires a deeper understanding and integration of theoretical frameworks and ethical considerations.