The application of artificial intelligence in collaborative learning has achieved more fruitful results, but it appears that one of the biggest obstacles in this field is the lack of big enough data set with good qualities. This work-in-progress paper describes a project which provides a way to solve this obstacle. This project is mainly aimed at college students and applicable to college reading courses (literary criticism, Literature Reading & Academic Writing, etc.). The collaborative reading activity mode proposed based on this project is adopted to provide a method for collecting students' highlighting, comments, responses and other data.