Engaging Artificial Intelligence (AI)-based chatbots in digital health: A systematic review

Area of Science:

• Health Informatics
• Artificial Intelligence
• Human-Computer Interaction

Background:

• The healthcare sector is increasingly adopting Artificial Intelligence (AI) solutions.
• AI-based chatbots are emerging as key tools to address health challenges.
• Deep learning architectures are driving advancements in AI healthcare applications.

Purpose of the Study:

• To systematically review existing research on AI-based chatbots in healthcare.
• To identify current applications and evaluation methods for AI chatbots.
• To address concerns regarding chatbot effectiveness and ethical implications.

Main Methods:

• Systematic literature review of 348 articles from eight major databases.
• Analysis focused on AI-based chatbots in the health domain.
• Identified four primary research areas: text quality, clinical efficacy, user engagement, and safety.

Main Results:

• AI-based chatbots are applied across four main research areas.
• A significant gap exists in the number of randomized controlled trials (RCTs).
• Limited use of theoretical frameworks hinders robust performance evaluation.

Conclusions:

• Further research is needed to optimize AI chatbot potential in healthcare.
• Addressing the lack of RCTs and theoretical frameworks is crucial for advancing chatbot development.
• Future directions should focus on enhancing clinical efficacy, user engagement, and safety while considering ethical implications.