Feedback Footprint
Augmenting teacher feedback with adaptive technology
This project develops an AI system to enhance teacher feedback on student writing. While current AI tools can identify basic writing issues, they often lack the nuanced, personalized guidance that skilled teachers provide.
The research combines teacher expertise with machine learning to create an AI writing assistant that understands effective feedback principles. By studying how expert teachers balance different aspects of feedback - from technical skills to student engagement - we're developing AI models that can generate contextual feedback matching individual teaching styles.
The system aims to help teachers provide consistent, high-quality feedback while preserving their unique approaches to student development. Key capabilities include adapting feedback to different skill levels and balancing technical and motivational guidance to support student growth as writers.
Related Publications
- Mah, C., Tan, M., Phalen, L. ., Sparks, A., & Demszky , D. From Sentence-Corrections to Deeper Dialogue: Qualitative Insights from Llm and Teacher Feedback on Student Writing. SSRN. https://doi.org/10.2139/ssrn.5213040
- Coelho, R., Levine, S., Abdi, D., Phalen, L., Harris, L., Demszky, D., & Lee, V. (2024). Middle School Students’ Engagement with Quantitative Data Representations of Fictional Texts. Proceedings of the 18th International Conference of the Learning Sciences-ICLS. https://repository.isls.org/bitstream/1/10709/1/ICLS2024_1398-1401.pdf
Awards
Two-Time Seed Grant Recipient, 2024 and 2025
Finalist for Data Prize, 2025