KAIST has now set a new benchmark in AI education with its virtual teaching assistant (VTA). Developed by Professor Yoonjae Choi from the Kim Jaechul Graduate School of AI and Professor Hwajeong Hong of the Department of Industrial Design, this assistant offers custom feedback even in large-class settings. It’s designed to support you with quick, precise answers—no matter the hour.
Jiwon Yang, a Ph.D. student at KAIST, summed up the experience: ‘At first, I didn’t have high expectations for the VTA, but it turned out to be extremely helpful—especially when I had sudden questions late at night, I could get immediate answers.’ Rolled out during the Fall 2024 semester in the ‘Programming for Artificial Intelligence’ course, the system engaged 477 master’s and Ph.D. students in a real-world test of its capabilities.
Unlike more general tools such as ChatGPT, this assistant is built on a Retrieval-Augmented Generation (RAG) system. In other words, it automatically sifts through extensive course materials—from lecture slides to video lectures—to provide you with responses that are both accurate and reliable. If you’ve ever struggled to find the right answer in a sea of information, this specialised approach could be just what you need.
Over a 14‑week term, more than half the students interacted with the VTA, generating 3,869 Q&A exchanges. Interestingly, students with less background in AI leaned on the tool more, suggesting that the assistant helped them explore new concepts without fear of getting things wrong.
The team’s survey results further underlined growing trust and satisfaction among students. Professor Choi pointed out, ‘This study shows that AI technology can be a practical ally for both students and instructors.’ To encourage broader adoption, the VTA’s source code is available on GitHub, inviting other institutions to adapt the system for their own classrooms.
