Institution: Maynooth University

Discipline:  Writing in Media

Authors: Adrian Kirwan

GenAI tool(s) used: ChatGPT

Situation / Context

The activities discussed below took place during the 2023-24 academic year at Maynooth University in a second-year module—Intersections of Science Writing and Media—which seeks to develop transferable writing skills for science students (n=64). The final six weeks of the module are built around the completion of a group podcast assignment which articulates a scientific topic to a public audience.

As the module is focused on writing as a rhetorical practice, students are well positioned to engage with the potential of GenAI focusing on Large Language Models (LLMs) as a writing technology. Initially, groups identified an appropriate topic and completed academic research.

This showcase focuses on the activities following the completion of this research when students produce a draft script for their podcast. It was at this point, over a two-week period that GenAI was introduced and utilised.

Task / Goal

The purpose of these activities was to train students to critically evaluate the use of GenAI in their writing. This necessitated the acquisition of two core competencies. Firstly, GenAI literacies—ways of doing and knowing, e.g. understanding the affordances and limitations inherent in LLM architecture, successful prompting, etc. Secondly, dispositions—needed to think critically about the issues that create and shape these technologies, including the ethical and legal dimensions of GenAI use, what contexts they are useful in and when it is more appropriate not to use GenAI (this can centre on concerns such as plagiarism; potential inaccuracies in output; impact on learning; environmental issues; or simply the fact that in many scenarios, using GenAI can be more time consuming than not). Drawing on Southworth et al., (2023), the learning outcomes were:

• Discuss how GenAI technologies work.
• Operate/prompt them effectively.
• Analyse their affordances and limitations.
• Integrate them into the writing process.
• Debate the broader social and ethical considerations of their use.

Actions / Implementation

The final assessment for this module was a group podcast. As outlined above, GenAI was the focus of the middle two weeks of this part of the course. The core teaching and learning activities were thus:

• Week one, lecture: Introduction to GenAI as a writing technology. This focused on three central learning outcomes: 1) describe the development and architecture of LLMs, 2) discuss the affordances and limitations of LLMs, 3) utilise LLMs. Following this lecture students took part in a tutorial session where they utilised GenAI to produce draft scripts.
• Tutorial. Three key prompting tasks were used: 1) brainstorming and organising, 2) producing text, 3) breaking down writing tasks to enhance GenAI output (see digital resources for worksheet).
• Week two, lecture: The ethics of GenAI. This focused on various ethical concerns raised by these technologies, including ecological, privacy, and copyright concerns; bias; and an overview and critique of current ethical frameworks.
• Tutorial. This consisted of critiquing GenAI output and editing this to produce their podcast script (see digital resources section).

Outcomes

The introduction of GenAI necessitated a multi-stage approach to the production of the podcast. While some elements remained the same, some changed. Preceding the use of GenAI, the module needed to introduce and discuss podcasting as a genre, and the role of science podcasting, to a public audience, in particular. Students needed to select an appropriate topic and research this, gathering academic sources which would allow them to critique, strengthen and correct potential misinformation from GenAI. Such fundamental knowledge is also a crucial factor in prompting LLMs. Following the production of a script with the assistance of GenAI (ChatGPT), students finally had to record and edit this for the assignment. The podcast script was assessed based on the following criteria: exigence (i.e. the pressing need), which was established, and resolution explored; audience was targeted; appropriate language, etc.; constraints, making use of the affordances of the genre while overcoming the limitations.

Reflections

The module proved extremely useful in exploring GenAI’s potential as a writing technology. The lectures provided a good opportunity to introduce key concepts, affordances, and limitations and begin to introduce core skills/literacies, such as prompting. These blended well with the tutorials, where students gained experience with generating and critiquing GenAI output.

One critical point of feedback from students was how time-consuming it was to produce usable text. Given the context—students had spent a lot of time up to this point thinking about writing and rhetoric—this not only signals the effectiveness of the module as a whole, but points to the need for strong writing instruction. Once students understand what proficient writing is, they are much more capable of critiquing and surpassing the textual output of GenAI.

Future iterations of this course will further explore the use of GenAI for research (compared to other tools) and summarisation (compared to humans).

Further Reading

Association for Writing Across the Curriculum. (2023). Artificial intelligence writing tools in writing across the curriculum.

Miao, F., Holmes, W., & United Nations Educational, Scientific and Cultural Organization. (2023). Guidance for generative AI in education and research. United Nations Educational, Scientific and Cultural Organization.

MLA-CCC. (n.d.). MLA-CCC joint task force on writing and AI.

Southworth, J., Migliaccio, K., Glover, J., Glover, J., Reed, D., McCarty, C., Brendemuhl, J., & Thomas, A. (2023). Developing a model for AI across the curriculum: Transforming the higher education landscape via innovation in AI literacy. Computers and Education: Artificial Intelligence, 4, 100127.

Digital Resources

Each of the two tutorials made use of a worksheet, these can be located here:

Tutorial one: Introduction to prompting Worksheet

Tutorial two: Editing GenAI workshop

Author Biography

Dr Adrian Kirwan (Critical Skills, Maynooth University) is an historian of science and technology. He has published in numerous national and international journals, as well as recently co-editing the eleventh volume of The Correspondence of John Tyndall (University of Pittsburgh Press, 2022). He also has a strong interest in the Scholarship of Teaching and Learning, particularly in the use of (and discourse surrounding) technologies in higher education. His most recent publication on this subject is ‘ChatGPT and university teaching, learning and assessment: some initial reflections on teaching academic integrity in the age of Large Language Models,’ Irish Educational Studies.