We currently use AI-related algorithms for three features. In general, we have taken a very careful approach to the use of AI. While we have not formally tested for bias, there is no known bias and, more importantly, the features are designed so that any bias is mitigated by keeping humans in the loop.
- Automatic scoring. This is a supervised machine learning approach to automatically grading comments and questions that students post, for thoughtfulness and effort. Perusall suggests quality scores for each comment and question posted by students, which instructors can optionally integrate into the student’s score for the assignment. Instructors always have full control over whether and how these scores are used, and can override them at any time.
- Suggesting discussion prompts. Instructors have access to a feature that allows them to highlight text and use generative AI to suggest a discussion prompt that they can present to students, to get their students thinking and discussing the selected text. Instructors have the ability to preview, edit, and accept or discard the suggested discussion prompt before it is presented to students.
- Confusion report. For each assignment, instructors have access to a “confusion report” that automatically summarizes the questions that students are asking about the assigned content. A generative AI algorithm is used to group student questions into related topic areas, summarize those topic areas, and pick out specific questions asked by students as exemplifying each topic. This allows instructors—particularly in large courses, where students write more questions than an instructor would have time to read—to use student questions to effectively plan out their class time and make connections between the questions students asked in the assigned reading (or video, podcast, etc) and what happens in class.
These features are provided by securely processing data in a private cloud environment in a way that adheres to our very strict privacy policy. For example, instructor/student data is never used by third-party providers to train their models. The main points of our privacy policy are:
- We only collect personal data when needed to provide the Perusall service. We intentionally avoid collecting data that does not directly serve a purpose for teaching and learning or internal product development.
- We only share personal data when it is necessary to provide the service. We share data only with your institution and with selected vetted, trusted, and established vendors that help us provide the service to you.
- We never sell your personal data. Perusall is not in the business of making money from your personal data (we make money when students purchase books or other materials through the platform or when your institution purchases special services from us).
Related to