Aarhus University Seal

Chatbot for exam simulation 

Short description

To prepare for an oral exam, students simulate the examination by engaging in dialogue with a chatbot. A student uploads their prepared material to the chatbot, which then asks follow-up questions based on the syllabus. The student answers the questions orally to their fellow students, who then evaluate the response. 

Motivation

  • The students are very nervous about the oral exam, including the format and the types of questions. 

  • It is necessary for them to understand how they can (and cannot) use chatbots in their academic work. 

Learning objectives

  • The students must be trained to take an oral exam in programming. Simulating an oral exam, using chatbot as an interactive examiner, helps them prepare for the exam and become familiar with some possible questions. 

  • The students need to learn how to write code. In this activity students reflect both on their code and aesthetic concepts, bridging code and theory. 

  • Working in groups, students practice collaborative interpretation of code and theory, enhancing their ability to discuss, justify, and refine ideas in dialogue with peers and the chatbot. 

Execution

Before class

Teachers’ preparation 

Preparation of prompt. For the teaching session, I prepare the prompt that the students will use for the activity. 

  • I prompt the chatbot and ask it to act as an examiner in the subject and pose questions about the students' work. 

  • The chatbot should ask the students to upload a piece of code, which it will then ask follow-up questions about. The follow-up questions should be both of a technical and conceptual nature. 

  • I upload the relevant curriculum material and ask the chatbot to only base its questions on the relevant chapters. 

Teachers’ preparation 

Preparation of assessment form. I prepare a form that the students will use to evaluate each other's oral responses to the exam questions. 

  • The form is structured according to the chapters in the curriculum — in this case, six chapters. 

  • For each chapter, the students must assess their peer’s response on a scale from 1 to 10. 

Students’ preparation  

Code Development. The six chapters from the course curriculum are covered in previous lectures. At the end of each chapter, the textbook includes an assignment designed to engage students with the chapter's key topics. These assignments are referred to as Minix. Each Minix typically consists of two components:  

  1. RunMe, which is usually a programming exercise 

  1. ReadMe, in which students articulate their aesthetic reflections on the assignment.  

In class

Plenary 

Introduction (5–10 min). I explain to the students that we are going to run an exam simulation based on the curriculum and their coding work. 

  • I explain that it is a pre-prompted chatbot that will act as the examiner, and I send them a link to access the chatbot. 

  • The students are divided into groups and choose one person at a time who will be examined by the chatbot. 

  • Each group receives an assessment form. 

  • I explain that the exercise is meant to be fun and that they should not be afraid of giving wrong answers. 

In groups 

First Round (10 minutes). The student who is about to be examined opens the chatbot link. 

  • The chatbot asks the student to upload the code from the RunMe exercise they have prepared for chapter 1. 

  • The chatbot then asks follow-up questions about the code. The questions are both technical and conceptual. 

  • The student answers the questions orally to the best of their ability, addressing their group members. 

  • Once the questions have been answered, the other group members discuss how well the answers were given. 

  • Finally, the other group members provide an assessment on a scale from 1 to 10. 

In groups 

Subsequent Rounds (50 min). The exercise is repeated until all six chapters have been covered. 

Plenary 

Wrap-up (10 min). I wrap up the exercise and talk with the students about how they experienced using the chatbot. 

  • I explain, that they can re-perform the activity at home to prepare for the exam. 

RESSOURCES FOR STUDENTS

SUPPORT FOR STUDENTS

  • Exercise guide in Brightspace 

  • Prompt for the students 

  • Coding in gitlab 

During the exercise, I circulate among the groups and answer questions where needed. This also gives me a sense of where their knowledge gaps are and what they struggle with. 

Reflections

Outcomes

  • The students had the opportunity to try out the exam situation and demystify it. They expressed that it made them feel less nervous about the exam. 

  • They gained insight into the types of questions that may be asked during the exam, helping them prepare more effectively. 

  • They also gained awareness of their own knowledge gaps and what they need to review for the exam. 

  • The students were required to justify the grade they assigned, which developed their reasoning and critical thinking skills. 

  • The students learned how chatbots behave and gained experience with the strengths and limitations of chatbots. 

  • Several students stated that they would use the prompt out of class when preparing for both teaching sessions and the exam. 

Challenges

  • The chatbot asked many good questions, but also some that did not make much sense or were phrased unnaturally. 

  • The chatbot does not always behave as desired. For example, it does not always ask questions but may instead perform a kind of review of the code uploaded by the student, which is an error. This particularly occur for students using a free version of a chatbot and therefore not having access to the most advanced version. 

  • The chatbot asked a lot of questions, which felt overwhelming to the students. It may therefore be a good idea to set a limit on how many questions it should ask per chapter, so students can cover all chapters.  

  • At times, the chatbot was unable to accurately measure the students’ maximum tolerance for difficulty, resulting in questions that were too challenging. This occasionally led to feelings of intimidation or anxiety among some students. 

Advices for other educators

  • Test your chatbot yourself before using it in teaching. 

  • After class, evaluate how the chatbot behaved and consider whether you should revise the prompt. 

  • As a teacher, you should be present during the session to provide assistance when needed. However, be careful not to participate too actively in group discussions, as it may diminish the spontaneity of the exercise. 

  • Consider when during the semester the activity should take place and whether it should be ahead of the exam. 

  • Make sure that the content of the exercise aligns with the exam. 

  • You may choose to provide students with criteria they can use when evaluating their peers. 

  • Consider whether students should give each other grades or just provide oral feedback. 

  • Consider whether to inform them about the exercise in advance. If they are told to prepare, there is a risk that some students may not show up.  

  • Students also benefit from sorting through the chatbot’s good and poor questions. 

  • Ask students to create a user account before class so they can access the best and most powerful version of the chatbot, or use Microsoft Copilot, which Aarhus University is licensed to use. 

  • Be brave and try the exercise in your teaching. Even if the chatbot does not deliver perfect answers. 

  • If you are using ChatGPT consider buying the Pro version to enable the upload of all textbook materials. This would allow the chatbot to have deeper contextual understanding and more accurate domain-specific knowledge, improving the relevance and precision of its questions and responses. 

  • Consider introducing a system for providing feedback on students’ answers, either automated or instructor-led. 


Basic information

Educator Vittoria Frau 
Faculty and department Institut for Kommunikation og Kultur – Afdeling for design og  
Degree programme Programming 
Level of study Bachelor level 2. semester 
Course/subject Digital Design - Aesthetic Programming 
Number of students 45
Extent Activity
Primary activity Practice, communication and discussion
Teaching format Small class teaching
Technology ChatGPT
Implementation Onsite

Contact

Please contact the editors at AU Educate if you have any questions about the content of the platform or if you need consultation on your teaching from one of the many skilled professionals at the Centre for Educational Development