Do People Engage Cognitively with AI? Impact of AI Assistance on Incidental Learning
Project Overview
The document explores the role of generative AI in education, particularly its impact on cognitive engagement during decision-making tasks. It reveals that while AI-generated recommendations can enhance immediate performance, they do not necessarily facilitate incidental learning. The research emphasizes that effective learning is linked to deeper cognitive engagement, which occurs when learners derive conclusions independently from AI explanations. Consequently, the findings advocate for the design of AI systems that encourage this kind of engagement rather than simply delivering recommendations. This approach aims to foster an educational environment where learners actively interact with AI, leading to more meaningful learning experiences and better retention of knowledge. Overall, the document underscores the potential of generative AI to transform educational practices by promoting deeper understanding and critical thinking among students.
Key Applications
AI decision support systems (e.g., Meal Assistant)
Context: Nutritional decision-making tasks for individuals seeking to understand meal choices better.
Implementation: Participants made decisions on meal choices based on AI recommendations and explanations in a series of experiments.
Outcomes: Participants who received only AI explanations showed greater learning gains compared to those who received recommendations.
Challenges: Merely providing recommendations with explanations does not ensure deep engagement or learning.
Implementation Barriers
Cognitive engagement barrier
Users may not engage deeply with AI-generated recommendations and explanations, leading to superficial processing of information.
Proposed Solutions: Design AI interactions that require users to synthesize information themselves rather than just providing recommendations.
Project Team
Krzysztof Z. Gajos
Researcher
Lena Mamykina
Researcher
Contact Information
For information about the paper, please contact the authors.
Authors: Krzysztof Z. Gajos, Lena Mamykina
Source Publication: View Original PaperLink opens in a new window
Project Contact: Dr. Jianhua Yang
LLM Model Version: gpt-4o-mini-2024-07-18
Analysis Provider: Openai