Decoding trust in large language models for healthcare in Saudi Arabia

Area of Science:

• Health Informatics
• Artificial Intelligence Ethics
• Human-Computer Interaction

Background:

• AI-powered conversational agents like ChatGPT are increasingly used for medical advice and decision support.
• Understanding user trust and adoption is critical for AI in healthcare.
• The Saudi Arabian context presents unique cultural factors influencing AI adoption.

Purpose of the Study:

• To investigate factors influencing user trust and decision-making with AI (ChatGPT) in Saudi Arabian healthcare.
• To analyze the relationships between system attributes (competence, reliability, transparency, security, trustworthiness, persuasiveness) and user satisfaction.
• To provide recommendations for enhancing AI trustworthiness in medical consultations.

Main Methods:

• Utilized constructs from trust in technology, Technology Acceptance Model (TAM), Health Belief Model (HBM), and usability frameworks.
• Employed Partial Least Squares Structural Equation Modeling (PLS-SEM) for data analysis.
• Tested 15 hypotheses regarding user trust and decision-making with AI in healthcare.

Main Results:

• Reliability, security, and transparency were identified as significant drivers of trust and decision-making.
• Ten out of 15 hypotheses were supported, emphasizing the importance of trust and satisfaction.
• Cultural influences on AI adoption in Saudi Arabia's healthcare sector were explored.

Conclusions:

• Trust and satisfaction are paramount for the adoption of AI tools like ChatGPT in healthcare.
• Enhancing reliability, security, and transparency is key to fostering user confidence in AI-assisted medical decision-making.
• The study provides insights for responsible AI deployment in Saudi Arabian healthcare settings.