This SBIR Phase I project is dedicated to building a scalable project-based learning education technology platform. The technology in this project is based on a chatbot that sustains ongoing, interdisciplinary conversations with each student. The platform is designed to provide middle school students with curriculum materials that expose them to authentic and engaging real-world problems that, over the course of weeks or months, come together to form a cohesive response to a project prompt. The goal of this project is to build and refine this new form of eLearning platform that has the potential to greatly improve engagement and learning outcomes, especially for low-income and minority students. This goal will be achieved by making the standards-aligned educational content meaningful and relevant to the interests of students and by catering the instructional materials to each individual student's skill level so they can move through the materials at their own pace towards mastery. To measure the impact of this technology on student outcomes, this project will involve state-of-the-art efficacy research in accordance with guidelines in the Every Student Succeeds Act (ESSA).This SBIR Phase I project is concerned with developing a forward-looking education platform built around a chatbot that sustains long term, goals-oriented conversations powered by artificial intelligence (AI) and based on a decision tree architecture. The highly-customized chatbot is designed to make project-based learning scalable by empowering students to choose and customize their own project theme from a set of options, and then providing conversational explanations and practice questions that illustrate how Common Core standards apply to the student?s project. This research undertaking will involve developing and training custom neural networks to interpret student responses and questions. It will also involve developing algorithms that correctly identify each student's level of reading comprehension and, independently, conceptual understanding. With an understanding of the student's level, another algorithm will be developed that selects the appropriate materials and resources at the student?s reading comprehension level. These algorithms will be refined through user testing and a preliminary study of the efficacy of this education intervention will be tested in a correlational study with statistical controls.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.
