Objective:

To explore patient perspectives on AI-generated communications in healthcare and the importance of empathy in messaging.

Approach:

• Study Design: Conducted 40 in-depth, semistructured interviews at a single academic health center.

Key Findings:

• Patients are amenable to AI-drafted messages if AI use is disclosed and clinicians retain oversight.
• Patients view portals as transactional and do not expect deeply personal messages for routine tasks.
• Preferences for message tone, length, and detail vary based on the clinical context.
• In lower-stakes situations, concise AI-generated responses are preferred; in higher-stakes situations, preferences are mixed.

Interpretation:

Limitations:

• The study reflects a specific snapshot in time and may not capture evolving patient perspectives as AI technology and social norms change.
• The sample was limited to a single academic health center, which may not represent broader patient populations.

Conclusion:

Sources:

• Owens et al Study