Patients' Perceptions Toward Human-Artificial Intelligence Interaction in Health Care: Experimental Study

Area of Science:

• Healthcare Technology
• Medical Informatics
• Patient Experience

Background:

• Artificial intelligence (AI) is poised to integrate into various clinical care aspects, including prognosis, diagnostics, and planning.
• Patient perception is a critical factor influencing the adoption and success of AI clinical applications.
• Ensuring patient safety and demonstrating benefits are paramount for AI integration in healthcare.

Purpose of the Study:

• To investigate patient perceptions of AI clinical applications' benefits, risks, and usability.
• To analyze how different healthcare service encounter scenarios influence these perceptions.

Main Methods:

• A 2x3 experimental design crossing health condition (acute/chronic) with clinical encounter type (AI substituting, AI augmenting, no AI).
• Online survey data collected from 634 individuals in the United States.
• Analysis of perceptions regarding privacy, trust, communication, transparency, liability, benefits, and intention to use.

Main Results:

• Encounter type and health condition significantly impacted perceptions of privacy, trust, communication, transparency, liability, benefits, and usage intention.
• No significant differences were observed in perceptions of performance risk and social biases across scenarios.

Conclusions:

• Incompatibility with instrumental, technical, ethical, or regulatory values can lead to AI rejection in healthcare.
• Risks associated with AI in diagnostics and treatment necessitate further research before widespread implementation.
• Regulatory bodies and healthcare institutions must collaborate to establish standards, guidelines, and continuous evaluation systems for AI safety and ethics.