Section A: Project Information
ClassInsight is an AI-powered tool developed by the University of Miami Law School’s AI & Law (MiLA) Lab to make classroom learning more interactive and efficient. By providing nearly instantaneous insights on student responses, it helps professors quickly gauge comprehension and adjust their teaching while giving students immediate, personalized feedback. Unlike grading tools, ClassInsight is designed to support learning, identifying misconceptions without assigning scores. It seamlessly integrates into classrooms by sending structured response forms directly to students’ emails, making it easy to use without requiring additional logins or software.
The platform runs on a full-stack architecture, with a Flask-based Python API handling AI analysis, Google OAuth for authentication, and SMTP for automated email delivery. The frontend, built with React, JSX, and SCSS, ensures a clean, intuitive interface for both students and professors. Using OpenAI’s NLP models, ClassInsight analyzes free-response answers based on accuracy, keywords, and comprehension. Future updates will allow professors to customize AI parameters, tailoring feedback based on strictness, conciseness, or specific learning goals.
By reducing the time and effort needed to assess student understanding, ClassInsight allows professors to focus more on teaching and less on manual evaluation. Students benefit from real-time feedback, helping them refine their responses and engage more actively in class. With a planned Fall 2025 launch, ClassInsight will continue evolving to support both live and asynchronous learning, making it a practical and scalable solution for modern education.
YouTube Link: https://youtu.be/egPLN5Z60P0
Section B: Participant Information
| Title | First Name | Last Name | Organisation/Institution | Faculty/Department/Unit | Phone Number | Contact Person / Team Leader | |
|---|---|---|---|---|---|---|---|
| Dr. | Or | Cohen-Sasson | University of Miami School of Law | AI & Law (MiLA) Lab Director | orcs@law.miami.edu | +972 54-448-0676 | |
| Ms. | Roni | Kennedy | University of Miami | College of Engineering | rtk45@miami.edu | +1 (856) 448-3731 | |
| Ms. | Talia | Berler | University of Miami | Master of Science in Data Science | tkb40@miami.edu | +1 (305) 710-2434 |
Section C: Project Details
The inspiration for ClassInsight originated from the desire to leverage Gen AI as a tool for educators to be more effective with class time, rather than a hindrance allowing students to skirt critical thinking exercises, which are intended to help them to better grasp concepts and provide their instructors insight into the status of the class’s learning as a whole. We realized there was no available application using AI that assisted educators in analyzing free response or short essay style questions in real-time, only more structured question styles such as multiple choice or fill in the blank. We hypothesized that an AI-based free response analysis tool would serve not only to improve pedagogical efficiency, but would help to integrate AI technologies into humanities courses in a beneficial and user-friendly manner. Educators in the humanities have largely been excluded from the benefits of educational technology, particularly as AI becomes increasingly integrated into learning and instruction. By creating an AI-based application that quantitatively and qualitatively compares students’ answers to a free response prompt with the instructors model response to the aforementioned prompt, and allows educators to personalize the AI agent’s analysis with various categories, we provide a simple, yet effective option for educators in the humanities to begin integrating educational technology in their teaching.
N/A (Stream 3)
Technical Implementation and Performance:
ClassInsight is a full-stack AI-driven educational tool designed for real-time response analysis and instructor insights. The platform integrates Flask (Python API), OpenAI NLP models, and Google OAuth to process student responses dynamically, providing instant feedback without storing data. The frontend, built with React, JSX, and SCSS, ensures a seamless user experience, while SMTP-based email automation delivers structured response forms directly to students.
Functional Architecture & Workflow:
ClassInsight’s Flask backend receives student responses, processes them using OpenAI’s GPT-4 NLP, and categorizes them based on accuracy, keyword relevance, and comprehension levels. Results are then visualized in a professor-facing dashboard powered by GPT-4. Instructors can customize assessment parameters, allowing them to fine-tune AI evaluation based on strictness, conciseness, or subject-specific needs. The tool operates without persistent data storage, ensuring lightweight processing and privacy compliance.
Development Timeline:
Phase 1 (Fall 2024): MVP trial run with GPT-4 integration, using real student responses for validation; Figma UI prototype.
Phase 2 (Spring 2025): Core backend and frontend functionality completion; AI model refinement for response analysis.
Phase 3 (Summer 2025): Usability testing with university partners; expansion of data visualization features.
Phase 4 (Fall 2025): Beta launch in live classrooms with iterative improvements based on real-world use.
Performance Metrics:
ClassInsight is evaluated on processing speed, accuracy, and user adoption. The system delivers real-time feedback within seconds, enabling professors to adjust teaching in the moment without disrupting lectures. Its NLP analysis aligns with instructor evaluation, ensuring reliable and meaningful feedback. Finally, user engagement rates will measure ClassInsight’s classroom impact, tracking how many students and professors actively use and benefit from the tool. By continuously refining performance, ClassInsight remains practical, scalable, and impactful in modern education.
ClassInsight offers a fresh approach to AI-assisted education by focusing on real-time analysis of free-response student answers—an area that most existing educational tools overlook. Unlike traditional AI grading systems that primarily handle multiple-choice or fill-in-the-blank questions, ClassInsight dynamically evaluates qualitative responses, providing immediate insights for both students and professors. The tool is not designed for grading but instead gives a holistic view of student comprehension, helping educators identify areas where students are struggling and adjust their instruction accordingly.
What makes ClassInsight stand out is its customizability and real-time adaptability. Professors receive live feedback, allowing them to refine their teaching strategies on the spot, rather than waiting until after an assignment is graded. The tool integrates seamlessly into existing workflows by sending structured response forms directly to students’ emails, eliminating the need for additional software downloads or platform logins. This streamlined approach makes it easy to use both in real-time and asynchronously, supporting in-class discussions, office hours, and even homework assignments.
Looking ahead, ClassInsight will continue to evolve with expanded customization options. Planned features include allowing professors additional ways to track learning progress overtime, customize AI analysis based on strictness level, conciseness, word limits, and other factors, and refine how responses are analyzed. This combination of automation and instructor control makes ClassInsight a versatile and innovative solution, particularly for humanities courses where AI-driven tools have been underutilized. By prioritizing usability, adaptability, and meaningful insights, ClassInsight offers a smarter, more effective way to enhance classroom engagement and learning outcomes.
ClassInsight is designed to be scalable and adaptable as user demand grows. The system is built using a modular AI framework, meaning new features and improvements can be integrated without disrupting existing functionality. As more educators adopt ClassInsight, its cloud-based processing and streamlined workflow ensure that it can handle increasing usage without compromising speed or accuracy.
To support long-term user engagement, ClassInsight is designed to fit seamlessly into existing classroom workflows. Instead of requiring additional software downloads, it integrates with email, allowing students and professors to interact with it effortlessly. Future updates will focus on customization options, such as enabling professors to adjust AI analysis criteria and track student progress over time. These enhancements will help the tool remain relevant as classroom needs evolve.
From an environmental sustainability perspective, ClassInsight reduces reliance on paper-based assessments and minimizes computing waste by avoiding unnecessary data storage. Since responses are processed and discarded after analysis, the system runs efficiently without accumulating excess digital clutter. By focusing on simplicity, efficiency, and adaptability, ClassInsight is built to grow with user needs while staying lightweight and accessible, ensuring long-term success in diverse educational settings.
ClassInsight helps make education more accessible and equitable by giving both students and professors real-time insights into learning comprehension. In large classrooms, it’s easy for some students to fall behind without professors realizing, especially in discussion-based courses where free-response answers are harder to evaluate at scale. By providing instant feedback on student understanding, ClassInsight ensures that every student, regardless of background or learning style, has a chance to engage with material and improve their comprehension. Rather than replacing human evaluation, it serves as a support tool that helps professors quickly identify where students are struggling and adjust their teaching accordingly.
One of the biggest social impacts of ClassInsight is its ability to bring AI-driven educational tools to subjects like the humanities, which have been largely left out of recent advancements in ed-tech. While multiple-choice assessments and structured responses are easy to analyze with existing tools, free-response answers—where deeper critical thinking happens—are often overlooked. By making AI useful without being intrusive, ClassInsight helps bridge this gap and ensures that students in all fields of study can benefit from technology-enhanced learning.
To measure impact, ClassInsight will track engagement levels among both students and professors, including how often the tool is used and whether it leads to better classroom discussions and understanding. Student surveys and professor feedback will help gauge whether the tool is improving comprehension and participation. Because every classroom is different, the platform will continue evolving based on user feedback, making sure it remains practical, adaptable, and genuinely useful in supporting learning rather than acting as a one-size-fits-all solution.
Personal Information Collection Statement (PICS):
1. The personal data collected in this form will be used for activity-organizing, record keeping and reporting only. The collected personal data will be purged within 6 years after the event.
2. Please note that it is obligatory to provide the personal data required.
3. Your personal data collected will be kept by the LTTC and will not be transferred to outside parties.
4. You have the right to request access to and correction of information held by us about you. If you wish to access or correct your personal data, please contact our staff at lttc@eduhk.hk.
5. The University’s Privacy Policy Statement can be access at https://www.eduhk.hk/en/privacy-policy.
- I have read and agree to the competition rules and privacy policy.