Background and general information
Oral examinations are a valued component of many courses at TU/e, offering direct interaction between students and examiners. However, they often lack structure, consistency, and objective quality control. This poses challenges for both assessment accuracy and the ability to provide meaningful feedback. The EXAMAI project addresses these issues by introducing AI-supported oral examinations. Specifically, it explores the use of language models (LMs) to facilitate dynamic, interactive assessments. Instead of relying solely on the judgment and questioning style of individual teachers, the project proposes a standardized system where students engage with an AI in real-time conversations. These language models will be trained on course-specific content and tailored to align with learning objectives. This approach aims to enhance fairness, provide consistent feedback, and allow students to rehearse or reflect on their performance. The pilot will take place within the course “Advanced Actuators Design” and is designed to support students’ understanding while reducing workload for instructors. EXAMAI reflects a broader shift toward integrating AI in education, not to replace educators, but to support more transparent and scalable assessment. Through this project, the team hopes to learn how AI tools can meaningfully enhance teaching and evaluation practices.
Goal or aim of the project
The goal of the EXAMAI project is to design, implement, and evaluate an AI-assisted oral examination system that improves the quality, consistency, and scalability of student assessment. This system will use language models trained on course materials to simulate interactive dialogues with students. During these assessments, the AI will ask course-relevant questions, respond to student input, and adapt its line of inquiry in real time. Students will receive detailed, structured feedback that supports deeper learning and helps them identify areas for improvement. For educators, the platform offers a standardized framework that can reduce subjectivity in grading and provide insight into student progress. It also enables opportunities for formative assessment, allowing students to practice with the system before formal evaluation. A key focus of the project is understanding how AI can support rather than replace the educator’s role, and how to ensure that the system respects values such as privacy, transparency, and inclusiveness. The pilot will generate practical insights into the technical, pedagogical, and ethical aspects of AI-assisted assessment, with the intention of refining the concept for broader application. Ultimately, EXAMAI aims to contribute to a more flexible, responsive, and student-centered learning environment.
For more information, please contact:
