Inform the uninformed: Improving Online Informed Consent Reading with an AI-Powered Chatbot

AI-powered conversational agents for educational engagement

Context: Applied in educational settings to facilitate interactions such as informed consent, knowledge retention, health information access, teamwork understanding, and qualitative data gathering. These agents are designed to enhance user engagement through conversational interfaces, enabling a more interactive learning experience.

Implementation: Utilizes AI technologies such as chatbots and conversational agents that guide users through various processes, adapt to user responses, and provide tailored information. These systems engage users in dialogue, whether for understanding consent forms, enhancing factual knowledge retention, accessing health information, identifying project teammates, or conducting surveys.

Outcomes: Improved user engagement and knowledge retention, enhanced access to information, better collaboration among students, and richer qualitative feedback compared to traditional methods. These tools foster a more interactive and personalized learning environment.

Challenges: Common challenges include ensuring accuracy and credibility of provided information, managing user trust in AI recommendations, refining the assessment of user preferences, and the complexity of personalizing interactions. Additionally, time-consuming interactions may deter engagement in some cases.

Technical

Chatbots may have limited understanding of natural language and complex queries, which could result in incorrect information and inadequate responses.

Proposed Solutions: Implementing a hybrid model that combines rule-based and AI components to enhance accuracy, along with continuous updates and training of the AI models to improve comprehension and response accuracy.

User Engagement

Participants may be deterred by the time and effort required to interact with the chatbot.

Proposed Solutions: Ensure that compensation reflects the time spent interacting with the chatbot to maintain participant motivation.

Trust Barrier

Users may have concerns about the credibility of information provided by AI chatbots.

Proposed Solutions: Implementing expert sourcing and transparent processes to ensure information reliability.

Implementation Barrier

Difficulty in accurately assessing user preferences and capabilities for effective team matching.

Proposed Solutions: Refining algorithms and incorporating user feedback to enhance matching accuracy.

Data Collection Barrier

Challenges in personalizing questions for conversational surveys.

Proposed Solutions: Utilizing adaptive learning techniques to tailor questions based on prior responses.

Project Contact: Dr. Jianhua Yang

LLM Model Version: gpt-4o-mini-2024-07-18

Analysis Provider: Openai