Social Life Simulation for Non-Cognitive Skills Learning
Project Overview
The document discusses the role of generative AI in education, particularly its application in enhancing non-cognitive skills through interactive storytelling platforms like SimuLife++. Central to this platform is the Sage Agent, an AI mentor that supports users by guiding them through decision-making and social interactions, thereby fostering self-awareness, motivation, and engagement in narrative experiences. User studies suggest positive outcomes from this AI assistance, although challenges persist, such as the effectiveness of AI feedback and issues related to narrative control. The integration of generative AI in educational tools is emphasized, with a focus on the potential for multiplayer interactions and the ethical implications of AI use in education. Furthermore, the document stresses the importance of rigorous evaluation to understand the long-term impacts of these technologies on learners. Overall, generative AI is portrayed as a promising avenue for enhancing educational experiences, particularly in developing vital non-cognitive skills through innovative interactive methods.
Key Applications
SimuLife++ with Sage Agent
Context: Educational platform for undergraduate students that combines interactive storytelling and AI mentorship to improve non-cognitive skills such as decision-making, empathy, and resilience. Learners engage in narratives, make decisions, and receive feedback from an AI mentor, creating a rich learning experience.
Implementation: The platform utilizes generative AI through the Sage Agent to create interactive storytelling experiences where users can make decisions, receive personalized feedback, and reflect on their choices. The platform is designed to foster engagement and motivation through narrative immersion.
Outcomes: Users reported increased motivation, improved self-perceptions, enhanced resilience and coping abilities, and higher levels of reflection on non-cognitive skills through engagement in narrative experiences.
Challenges: Users expressed concerns regarding the AI's feedback being too vague or lengthy, which sometimes disrupted the narrative flow. There is also a need for more context-specific guidance and user control, and reliance on the OpenAI API may affect the quality of feedback.
Implementation Barriers
Feedback Clarity
The Sage Agent's comments were sometimes perceived as too abstract or lengthy, making it difficult for users to apply them effectively.
Proposed Solutions: Shortening the feedback and making it more context-specific could enhance its applicability and effectiveness.
User Engagement and Ethical Barrier
Participants felt that the predefined storylines limited their sense of agency and control over the narrative. There are also concerns about users role-playing as 'bad people' and making harmful choices, alongside the need for privacy and safety protocols.
Proposed Solutions: Improving user control over story progression and incorporating more dynamic character relationships could enhance engagement. Additionally, implement clear protocols for handling sensitive information and ensure ethical storytelling practices.
Research Barrier
Lack of longitudinal studies to assess long-term impacts of the AI system on non-cognitive skills.
Proposed Solutions: Future research should include longitudinal study designs and robust evaluation techniques.
Technical Barrier
The reliance on a general OpenAI API instead of a tailored model may affect narrative and conversational quality.
Proposed Solutions: Development and integration of a fine-tuned model for better user experience.
Project Team
Zihan Yan
Researcher
Yaohong Xiang
Researcher
Yun Huang
Researcher
Contact Information
For information about the paper, please contact the authors.
Authors: Zihan Yan, Yaohong Xiang, Yun Huang
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