Inspiration
Many students memorize concepts without knowing whether they truly understand them. We were inspired by the "teach-back" learning method, where explaining a concept in your own words helps reveal what you actually understand. This led us to create an AI platform that can listen to a student's explanation and identify areas where they need improvement.
What it does
TeachBack-AI allows students to explain a concept in their own words to an AI. It analyzes the explanation, identifies knowledge gaps and misconceptions, and provides personalized feedback. Based on the student's weak areas, it also generates targeted quizzes to help strengthen their understanding.
How we built it
We built TeachBack-AI as an interactive AI-powered learning platform. The system takes the student's explanation as input and uses AI to evaluate the content, identify missing or incorrect information, and generate understandable feedback. A quiz module then creates questions based on the identified learning gaps.
Challenges we ran into
One of the main challenges was making the AI accurately understand different ways students explain the same concept. We also had to ensure that the feedback was clear and constructive instead of simply marking an explanation as right or wrong. Designing a useful flow from explanation to feedback and quiz was another challenge.
Accomplishments that we're proud of
We are proud of creating a learning approach where students actively demonstrate their understanding instead of only consuming information. TeachBack-AI can turn a student's own explanation into personalized feedback and practice questions, making the learning process more interactive and meaningful.
What we learned
We learned how AI can be used for personalized education and how prompt design can influence the quality of AI-generated feedback. We also gained experience in designing an AI-based learning workflow and understanding how knowledge gaps can be identified from natural-language explanations.
What's next for TeachBack-AI
We plan to improve the accuracy of concept evaluation and make the feedback more personalized. Future versions could include progress tracking, difficulty-based quizzes, voice-based explanations, learning analytics, and support for multiple subjects and languages.
Log in or sign up for Devpost to join the conversation.