Abstract

Objective: This study aimed to assess the prevalence, patterns, and behavioral impact of artificial intelligence (AI) chatbot use for symptom assessment and laboratory results interpretation among adults in Saudi Arabia, in addition to identifying predictors of potentially unsafe behaviors.
Methods: A cross-sectional online survey was conducted in Saudi Arabia in December 2025, distributed via social media and in-person sharing. Adults aged ≥18 years residing in Saudi Arabia, who reported using AI-chatbots for the last 3 months, were eligible. Potentially unsafe behavior was defined as reporting ≥1 of the following after AI advice: delaying/avoiding medical care, changing prescribed medication without consultation, relying on AI advice without medical verification, or self-reported harm.
Results: Out of 2,011 participants, 1,398 (69.5%) reported using AI for symptom assessment, and 719 (35.8%) for laboratory results interpretation. AI users were younger (25.3 vs. 30.7; p < 0.001). Most users reported infrequent use (64.9% less than monthly). The most common cause for using AI was rapid access to information (78.4%) and understanding possible causes (60.4%). Among AI users, 35.3% met the high-risk criteria. High frequency use and high trust in AI were independent predictors of high-risk behavior.
Conclusion: AI chatbots were commonly used for symptoms and laboratory interpretation in Saudi Arabia, with over one-third of users reporting at least one potentially unsafe AI behavior. Higher trust and more frequent use were the key predictors.

Key words: Artificial intelligence, chatbots, AI, health behavior, Saudi Arabia