Novobo: Supporting Teachers' Peer Learning of Instructional Gestures by Teaching a Mentee AI-Agent Together
Project Overview
The document explores the application of generative AI in education, particularly highlighting Novobo, a teachable AI agent that enhances teachers' peer learning of instructional gestures. By fostering collaborative learning, Novobo enables educators to externalize, exchange, and internalize tacit knowledge about nonverbal communication, which is crucial for effective teaching. The study demonstrates that Novobo significantly reduces social pressure during peer feedback sessions while promoting a deeper understanding of instructional gestures through a structured interaction model. Furthermore, the document underscores the broader applications of generative AI in educational settings, showcasing tools that facilitate interaction, feedback, and analysis of teaching methods. This technology plays a vital role in understanding teacher behaviors, enhancing peer feedback, and improving student engagement, ultimately contributing to a more dynamic and responsive learning environment. Overall, the findings indicate that generative AI can significantly enrich the educational experience by supporting both teacher development and student learning engagement.
Key Applications
AI-powered tools for nonverbal behavior analysis and instructional gesture training.
Context: K-12 and higher education, focusing on teacher training programs and professional development. Teachers interact with AI systems to analyze video recordings of their teaching and receive feedback on both their instructional gestures and nonverbal behavior.
Implementation: Utilization of AI algorithms, including graph convolutional networks and teachable AI agents, to analyze video recordings and facilitate peer learning through a structured interaction process. This includes posing questions, commentary, demonstration, and explanation of instructional gestures.
Outcomes: Enhanced understanding of effective teaching behaviors, increased awareness of teaching strategies, and improved reflective practices among teachers, leading to collaborative learning and professional development.
Challenges: Initial skepticism towards AI-generated feedback, data privacy concerns, social pressure related to peer evaluations, and the need for contextual adaptability and training in using these tools effectively.
Video annotation tools and teachable agents for reflective practice and personalized learning.
Context: Higher education, particularly in teacher education programs, mathematics, and science courses. Teachers and students utilize video annotation tools and interact with AI-driven teachable agents to review and reflect on teaching practices and enhance learning engagement.
Implementation: Integration of video annotation tools and AI-driven teachable agents that assist in learning through interaction, enabling users to reflect on their teaching and learning practices effectively.
Outcomes: Increased student engagement, enhanced personalized learning experiences, and fostered professional development through deeper reflection on teaching strategies and student engagement.
Challenges: Technology accessibility, varying levels of student adaptability, and scalability of the technology.
Chatbots for data collection and feedback mechanisms.
Context: Higher education, focusing on student engagement through the collection of user self-reported data and feedback. Chatbots interact with students to gather insights and improve the educational experience.
Implementation: Development of chatbots leveraging large language models to interact with students, collecting self-reported data and feedback to enhance data collection processes.
Outcomes: Improved data collection processes, enhanced student feedback mechanisms, and fostered a more interactive learning environment.
Challenges: Ensuring user trust, the reliability of chatbot responses, and addressing privacy concerns.
Implementation Barriers
Social Barrier
Teachers may feel uncomfortable providing critical feedback to peers due to social dynamics.
Proposed Solutions: Positioning the AI as a mentee reduces social pressure and encourages more honest feedback.
Trust Barrier
Skepticism towards AI-generated suggestions and concerns about reliability and accuracy may lead to hesitance in using AI tools.
Proposed Solutions: Providing well-referenced theoretical knowledge alongside AI suggestions to increase trust, enhance user education on AI capabilities and limitations, and ensure clear communication regarding AI functionalities.
Technical Barrier
Data privacy concerns regarding the use of student and teacher data in AI applications.
Proposed Solutions: Implement robust data protection measures and transparency in data usage policies.
Accessibility Barrier
Not all educators and students have equal access to the technology required for generative AI applications.
Proposed Solutions: Invest in infrastructure improvements and provide training for all users.
Scalability Barrier
Challenges in scaling AI solutions across diverse educational contexts.
Proposed Solutions: Develop flexible AI solutions that can be tailored to different teaching environments and learning scenarios.
Project Team
Jiaqi Jiang
Researcher
Kexin Huang
Researcher
Roberto Martinez-Maldonado
Researcher
Huan Zeng
Researcher
Duo Gong
Researcher
Pengcheng An
Researcher
Contact Information
For information about the paper, please contact the authors.
Authors: Jiaqi Jiang, Kexin Huang, Roberto Martinez-Maldonado, Huan Zeng, Duo Gong, Pengcheng An
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