Large language models for surgical informed consent: an ethical perspective on simulated empathy

Area of Science:

• Medical Ethics
• Artificial Intelligence in Healthcare
• Surgical Patient Care

Background:

• Informed consent in surgery necessitates accurate information and empathic engagement to build trust.
• Large language models (LLMs) present opportunities to enhance informed consent through information retrieval and simulated emotional responsiveness.
• Ethical concerns regarding patient autonomy, trust, and transparency arise with simulated empathy in LLMs.

Purpose of the Study:

• To examine challenges in surgical informed consent.
• To evaluate the benefits and limitations of LLMs in this context.
• To explore the ethical implications of simulated empathy in LLMs for informed consent.

Main Methods:

• Distinguishing between active empathy (simulating emotion) and passive empathy (recognizing distress cues).
• Analyzing the potential of LLMs to identify and signal patient anxiety.
• Proposing a framework for ethical LLM integration in informed consent.

Main Results:

• Active empathy simulation by LLMs risks misleading patients and undermining trust.
• Passive empathy, focusing on recognizing and signaling distress, is ethically preferable.
• LLMs can analyze patient cues to alert healthcare providers to anxiety.

Conclusions:

• LLMs should be utilized for passive empathy, recognizing patient distress signals.
• This approach leverages LLM analytical strengths while preserving authentic human empathy.
• Ethical LLM use in informed consent supports, rather than replaces, the relational trust vital for patient-centered care.