Institution: Dublin City University

Discipline: Interdisciplinary Electronic Engineering and the School of Law & Government

Author:  Leah Ridgway

GenAI tool(s) used: ChatGPT

 

Situation / Context 

BSc Global Challenges at Dublin City University is an interdisciplinary course run by the School of Electronic Engineering and Law & Government. The activity in this case study was run in a seminar setting with first-year students in a small class of seven students. Within the “Shaping Global Leaders” module, students must complete a series of reflective portfolio pieces using the Driscoll Reflective Framework (2007). The example here is around reflecting on teamwork, resulting in this activity being suitable for any discipline where professional development, teamwork and reflective learning are part of the curriculum.

Task / Goal 

My own pedagogy has been influenced by my disciplinary background as an engineer and my own constructivist views, resulting in my desire to empower students to engage with and explore new technologies and tools and not shy away or ban their usage. I wanted students to see the advantages and the limitations of GenAI, but more importantly, to understand the design process of using prompts and reviewing and refining responses as part of an engineering design cycle rather than treat the technology as something that is taboo to discuss in an educational space. Collectively, as a society, we adapted to the introduction and adoption of calculators, so how do we learn to work with this newer tool?.

It was key for students to develop their own awareness of the best use cases for the technology. In this module, I presented a “good” and “bad” example of using the technology to enable a class discussion on the technology’s merits and limitations.

In the “good” example, developing case studies is an appropriate use of GenAI tools as the prompt design is the vital part of the content; our role here is to be the editor for a piece of fiction which needs to fit a brief that we provide. This example demonstrates to students how using critical reading and editing skills is an important part of the work – the human-supervised part. The GenAI can create what it wants, whereas we are responsible for the quality assurance. The key distinction lies in the fact that we remain actively involved in the process rather than outsourcing it and relinquishing our responsibility.

The contrast of a “bad” example of GenAI was provided by asking students to assess a reflective piece that had been fully outsourced with no editing, by providing the coursework instructions as the prompt. The aim was to get students to discuss the work and assess it against a rubric so they could get a lecturer’s perspective on why these pieces are not academically strong and apply this understanding to their own work in the future.

Actions / Implementation 

I used GenAI to produce group work case studies which discuss different types of group dynamics that may occur and sample student submissions reflecting on experiences of group working. These were discussed in class, and then students provided feedback and graded the sample reflections against a rubric.

• The “good” usage example:

The prompts used to generate the case studies, and the refinement process were discussed as an adapted model of engineering design. Here, the design is prompt design; the prototype is the GenAI output; this is evaluated and then improved upon if needed. The prompt design was discussed in the context of thinking of the required outputs and including these within the prompts. For example, “the group in the case study should have five students”. The following is an example of one case study design.

Prompt:

Write a case study which discusses a team of five university students working on a project together where there is one member of the team who does not contribute. The project is a semester-long group project, and the students are from different programmes.

The response produced a case study where a student named Emily was highlighted as “the team’s weakest link”, giving a list of her “shortcomings” and the consequences for the team. We discussed that the framing of the output was blaming one student without exploring the possible reasons for their behaviour. The chat was extended with the following:

Prompt:

Why did Emily act this way?

The response produced a list of examples, one of which was incorporated into the next prompt. This was selected because as part of the design, we wanted to discuss solutions to create groups that work.

Prompt:

Write a case study which discusses a team of five university students working on a project together where there is one member of the team (named Emily) who does not contribute. The project is a semester-long group project, and the students are from different programmes. Emily finds it difficult to attend group meetings and complete work on time because she has a job outside of college. Also, include hypothetical emotional responses from the rest of the group to the challenges.

The result was discussed as a case study in class for the actions the group members could take to ensure the overall success of their project and work with Emily to find solutions.

• The “bad” usage example:

The second generated piece was designed to be a “bad” usage example. This was an un-engineered response to supplying the coursework briefing, in which students were asked to reflect on their experiences of starting university.

The students were asked to read the work, which was presented as anonymised student work, and grade it against the rubric that would be applied to their own work. Initially, the work seemed good, but when scrutinised against the rubric categories in more detail, the weaknesses became clearer. The rubric required clear, specific examples and actions, which were not evident in the sample work. After this exercise, the work was unmasked as being produced by GenAI, followed by a discussion on the relevant distinguishing features which would identify it as such. The fundamental nature of GenAI as a language model and not a replacement for personal experience, reflection and analysis was discussed.

Our class discussion focused on the outputs of the reflective sample work being generic and inauthentic when analysed against the rubric and not satisfying the requirements for the best grades. This challenged students’ views as to the realities of the technology and how to recognise its limitations.

The two different document types were used to hold a group discussion on the strengths and weaknesses of the sample work and of using GenAI tools for more speculative writing and editing.

 

Outcomes 

Students were empowered to discuss GenAI and the appropriateness of its usage in different contexts by providing a “good” and “bad” example. During the session, the group co-created an updated rubric based on their grading of sample work created by GenAI, which would be applied to their own work. This emphasised the importance of authenticity and specific examples.

Reflections 

The process employed made it easier for students to discuss the work’s strengths and weaknesses because it was created using GenAI rather than anonymised student work.

From my own perspective, I learned that first-year students could find it hard to discern the quality of GenAI-produced outputs initially. The use of the rubric grading and related discussion facilitated students’ development of more sophisticated analysis skills related to their writing.

Further Reading 

Driscoll, J. (Ed.). (2007). Practising clinical supervision: A reflective approach for healthcare professionals. Edinburgh: Elsevier.

Author Biographies 

Dr Leah Ridgway is an Assistant Professor in the School of Electronic Engineering at Dublin City University and is part of the Centre for the Advancement of STEM Teaching and Learning (CASTeL). They are a Senior Fellow of Advance HE and have more than a decade in teaching within engineering. Leah’s research interests are in engineering education within the university sector using a qualitative and quantitative approach to investigate how students learn.