Problem and Need on campus: The SUNY General Education Framework (SUNY GE, 2025) includes AI in Information Literacy as a required core competency that “students will … demonstrate an understanding of the ethical dimensions of information use, creation, and dissemination, whether from traditional sources or emerging technologies, such as artificial intelligence.” More specifically, the guidance indicates that “Students need to acquire information literacy appropriate to the demands of the 21st century, which includes applying all three learning outcomes of this core competency to various sources of information, including emerging technologies, such as artificial intelligence.”

At SUNY Brockport, despite the rapid growth and easy accessibility of AI tools, there is no policy regulating their use. However, the university librarians have developed a library site offering AI-related resources. Faculty have the autonomy to decide the extent to which their students are allowed to use AI, ranging from encouraging its use with reflection and appropriate citations, to imposing restrictions, and prohibiting its use entirely. It is urgent to provide students with opportunities to use AI, under appropriate guidance.

Effective Practice: Feeling a sense of urgency, Dr. Ning Yu and Dr. Sandeep Mitra mentored their Computer Science (CS) students in the inaugural Association of Computer Machinery Special Interest Group in Artificial Intelligence (ACM SIGAI) Student Chapter. They collaborated with Dr. Jie Zhang, a professor of Special Education, following a “learning by doing” model, to sharpen students’ computer science skills, particularly in AI.

A flow chart was first created to outline the project structure. Dr. Zhang selected the Autism Module in her Introduction to Special Education course and its assessment, Autism homework, to be piloted in fall 2024. She collaborated with the team to outline how her students would engage with the assessment in two attempts and supplied the team with all relevant teaching materials, the questions, rubric, and answer key for the pilot.

In the first attempt, students answered eight short-answer questions on autism. After their submission, AI provided immediate and individualized feedback tailored to the student’s answers and graded students’ responses based on a predefined grading rubric. Then, for the second attempt, students were asked to revise their answers to the original eight questions for improvement. They also self-evaluated their own performance based on the improved answers and responded to two additional reflective questions about their experience using AI in the process. After students’ second submission, Dr. Zhang reviewed the students’ improved work and provided further feedback and grades.

Key features of this application include:
• Device-Independent and Offline Functionality: The application works on any device or operating system and can operate offline, ensuring availability for all users regardless of location.
• Speech-to-Text Function: Enables students to answer questions verbally during assessments, offering a flexible and accessible way for students to complete assessments that may otherwise be limited by traditional methods.
• Anonymous Submissions: To ensure student privacy, the application facilitates anonymous submissions. For grading purposes, submissions require the last four digits of each student’s school ID, which can be verified by the course instructor only. This anonymization creates a comfortable environment for students to participate in assessments without fear of reidentification.
• AI-Driven Feedback System: Powered by prompt engineering, this system compares students' responses to the provided teaching materials and answer key, delivering customized, constructive, and step-by-step feedback that builds confidence and encourages deeper learning. This two-step approach first provides initial positive feedback on what the student did well, then identifies areas for improvement while directing them to relevant sections of the material for further study. It is especially effective in practice exams or early assessments, allowing students to self-assess and focus on targeted improvements.
• Automated Grading: The application automates the assessment process, significantly reducing the time required to complete assessments and enhancing efficiency for both students and educators. The AI-powered automated grading function is based on the given instruction materials, rubric, and answer key, ensuring that responses are graded accurately and consistently. This feature reduces the workload for teachers, allowing them to focus on providing personalized feedback to students. The automation minimizes human error and enhances grading efficiency, making the assessment process more manageable and effective, even in virtual classrooms.

Why it was effective: The proposed project aims to deepen student learning by integrating AI technologies into teaching. It addresses issues of assessment accessibility and inclusivity, and gauges students’ understanding through AI-empowered educational assessment. More specifically, it utilizes immediate, targeted, and personalized AI feedback and automated grading. By doing so, the study tutoring experience becomes more supportive and equitable. Furthermore, this project helps lower the assessment stakes and focus more on the learning process.

Reflections and Implications: This AI-driven two-step feedback system, empowered by prompt engineering and embedded with automated grading, was designed to enhance accessibility and inclusivity in educational assessments. Through its automated feedback and assessment capabilities, the application engages students in personalized learning, demonstrating their knowledge, and fostering a more inclusive educational environment. Rather than serving as a replacement for educators, the AI complements educator support and assists students in monitoring their learning through self-assessment and automated feedback. By identifying and addressing knowledge gaps, the application empowers learners to take control of their education. Personalized positive feedback enhances students’ understanding and overall learning experience, as Hattie and Timperley (2007) stated, positive feedback can increase the likelihood that students will continue attempting an activity and maintain a higher level of interest. This design allows educators to focus on providing meaningful insights and fostering valuable interactions, while the technical aspects of assessment are efficiently managed by the application.

Future Directions: Future updates of this application will include secure logins for both teacher and student accounts, enhancing management and oversight while ensuring the privacy and security of user data. The team will also explore expanding the implementation of this application to additional disciplines, such as biology and mechanical engineering, to further support diverse educational needs across a broader range of content areas.

Fonseca-Llorca, L., & Davies, L. (2024, November 15). EduAlly pilot results: Insights, improvements, and future vision. Brockport, NY, November 15, 2024. The AI-OER Faculty Learning Community Meeting.
Fonseca-Llorca, L. E., Zhang, J., Davies, L., Mitra, S., & Yu, N. (In review). Developing an AI-empowered test-taking application: Enhancing accessibility, confidentiality, productivity, and student learning in education. Curriculum and Teaching Dialogue.
Hattie, J., & Timperley, H. (2007). The power of feedback. Review of Educational Research, 77(1), 81–112. doi: https://doi.org/10.3102/003465430298487
Krstić, L., Aleksić, V., & Krstić, M. (2022). Artificial intelligence in education: A review. The 9th International Scientific Conference Technics and Informatics in Education – TIE 2022, 223– 228. doi: 10.46793/TIE22.223K.
State University of New York. (2025, January 8). SUNY General Education Framework (SUNY GE). https://system.suny.edu/academic-affairs/acaproplan/general-education/suny-ge/
Zhang, J. (2004). Introduction to Special Education Course Materials.
Zhang, J., Yu, N., Li, W., Mitra, S., Begany, G., Zhang, S., Rich, A., & Hung, J. (2025). Enhancing learning across disciplines through AI-empowered interactive education. Washington, DC, January 22–24, 2025. The American Association of Colleges and Universities (AAC&U) Annual Meeting: Reaffirming Higher Education’s Public Purpose.

Student Feedback - Positive Feedback
• Helpful & accurate AI feedback
• Quick insights for improvement
• Streamlined submission process
• Overall Sentiment: Concept appreciated, with suggestions for improvements in future use

Student Feedback - Suggestions
• Feedback Clarity: "Elaborate" feedback was too vague
• Technical Issues: Feedback only applies to answers saved in the current session. Saved answers from previous session were not visible after exiting the exam.

The application was created. The student team can establish more assessment modules in the application. No other hardware or software is needed to implement this effective practice.
No costs are associated with this practice. This project is part of the IITG 2024-25 grant. However, there were no costs associated with the pilot conducted in the fall 2024 semester, as we did not receive the funds until January 2025.