How Following Medical Artificial Intelligence Advice Can Mitigate Malpractice Liability: Cross-National Insights from...

Area of Science:

• Medical ethics
• Artificial intelligence in healthcare
• Legal implications of AI

Background:

• Artificial intelligence (AI) is increasingly used in clinical decision-making, but its recommendations can differ from standard medical practices.
• Malpractice concerns may influence physician adoption of AI, yet prior US research indicated jurors hold physicians liable for rejecting AI advice.
• The impact of AI recommendations on perceived physician liability in legal systems relying on court-appointed experts remains unexplored.

Purpose of the Study:

• To investigate how physicians and laypeople in expert-based and lay-juror legal systems evaluate physician adherence to AI recommendations.
• To assess perceptions when AI advice deviates from standard medical care.
• To compare these evaluations across the United States and Germany.

Main Methods:

• A randomized vignette study with a 2x2 factorial design was employed.
• Participants evaluated fictional physicians' decisions to accept or reject AI recommendations (standard vs. nonstandard care).
• Online surveys were conducted with nationally representative samples of U.S. adults, German adults, and German physicians.

Main Results:

• Physicians accepting standard-care AI recommendations were rated as more reasonable than those rejecting them across all participant groups (U.S. laypeople, German physicians, German laypeople).
• Accepting or rejecting AI recommendations for nonstandard care resulted in statistically equivalent reasonableness ratings.
• These findings held true in both U.S. lay-juror and German expert-based legal system contexts.

Conclusions:

• Accepting standard-care AI advice is perceived as more reasonable by lay jurors and court-appointed experts in both the U.S. and Germany.
• The decision to accept or reject nonstandard-care AI advice did not significantly alter perceived reasonableness.
• Malpractice liability systems may not inherently hinder the integration of AI in precision medicine.