Stanford University researchers unveil Tutor CoPilot, an AI system designed to assist tutors in enhancing virtual maths lessons by fostering a deeper understanding in students.
Stanford University researchers have developed an innovative AI system called Tutor CoPilot, which is built on OpenAI’s GPT-4 language model. This system has been integrated into a virtual learning platform, FEV Tutor, which facilitates online interactions between students and tutors. The primary aim of Tutor CoPilot is to assist tutors in effectively guiding students toward understanding and correcting their errors during virtual maths lessons.
The process of development for Tutor CoPilot involved training GPT-4 on a substantial database of 700 actual tutoring sessions. These sessions featured experienced educators working with students from first to fifth grade. The educators focused on recognising student errors and promoting understanding of broader mathematical concepts. The model was designed to generate suggestions that tutors can adapt to individualised, online teaching scenarios.
Rose Wang, a PhD student at Stanford University and a key contributor to this project, expressed enthusiasm for the potential of AI-assisted learning. “I’m really excited about the future of human-AI collaboration systems,” commented Wang regarding the project, whose research findings have been published on arXiv, although they have yet to undergo peer review. “I think this technology is a huge enabler, but only if it’s designed well.”
The objectives of Tutor CoPilot are clear: the system is not intended to replace teaching or provide direct educational content to students. Instead, it functions as a support tool for tutors by suggesting strategies to better steer students towards discovering the correct solutions on their own. It’s a tool that emphasises fostering a deeper understanding in learners by suggesting pedagogical techniques to prompt critical thinking. For example, the AI might recommend tutors to ask students how they arrived at an answer, or propose alternative questions that encourage exploration of different problem-solving approaches.
The introduction of Tutor CoPilot underscores an ongoing interest in harnessing AI technology to complement human teaching methods. As education continues to evolve in the digital era, tools like Tutor CoPilot represent promising advances in personalising online learning experiences and enhancing the effectiveness of virtual tutoring sessions.
Source: Noah Wire Services
