Breaking the Stalemate: Integrating Generative-AI and Media Literacy into Higher Education Pedagogy

The Program

Apr 25

Written By

Ryan D'Souza, Ph.D.
Arts, Design & Communication

Project Overview

I am interested in addressing the stalemate in the classroom between faculty and students concerning generative-AI tools. The notion of a stalemate is borrowed from chess, wherein neither side can progress or win, leading to a draw; however, in the classroom, while there are no immediate winners, there is a clear loser – the debt-incurring student. Following reports from the World Economic Forum, generative-AI tools are now part of the desirable “soft skills” employers look for (2023), as well as the political acumen for citizens living in a media-dense world (2024).

In the context of generative-AI, the stalemate is also between orientations toward technologies. Whereas the industry is technophilic, integrating mass-produced technologies as they become available (such as the desktop computer in the early 90s), higher education is technophobic, often deliberating the use of technologies within narrow definitions of academic integrity (such as policies for generative-AI tools). However, if a pipeline is assumed, such that education links students to the ballot box and the workforce, the consequences of such divergent orientations are a media illiterate populace. We are increasingly finding ourselves in situations of mass availability of technologies juxtaposed against mass illiteracy.

In this regard, I approach generative-AI tools, first and foremost, as media. It is helpful to think about generative-AI within the sphere of media literacy because most universities already dedicate resources toward such programs. Then, if Marshall McLuhan is right, “The medium is the message,” the concern with generative-AI tools is one of inter-generational communication differences. In the sense, the medium is reorganizing consciousness, relationships, and society.

In the midst of such changes, faculty must adopt, learn, and teach. Therefore, we must develop ways to better integrate new technologies and their communicative logics in research, teaching, and service. If not, both faculty and students will find themselves outside the sphere of social changes.

Planning Process

For the project, I centralized ChatGPT as a language-modelling tool for use in the writing-intensive course COM234W: Persuasion. I chose the course for convenience – it is offered every Spring, hence providing the conditions to execute the project regularly, and it involves consistent opportunities to integrate generative-AI tools into the modules. Moreover, taking the course as a writing-intensive one, using generative-AI with a focus on prompt writing addresses the requirements of course instruction at a micro-level, i.e., writing to initiate the analysis.

In the development of the course module, I utilized Bloom’s taxonomy to organize students into domains, i.e., cognitive, affective, and psychomotor. These domains allow me to formulate communication strategies for any shortcomings. For instance, several students expressed negative attitudes (affective) toward using ChatGPT for assignments, suggesting that using the platform was “cheating,” “dishonest,” or violating academic integrity. In such a case, I explained to students that it will be time-consuming to manually process the data of 5000 words. Moreover, showing students similar platforms, like SPSS, helped to reduce their negative affect.

In most cases, students demonstrated the required skills (psychomotor) to use ChatGPT. However, several students struggled to exhibit the knowledge (cognitive) required to navigate and troubleshoot the platform. I addressed these shortcomings with two other pedagogy models – critical communication pedagogy (CCP) and technological pedagogical content knowledge (TPACK).

One of the foundational understandings of CCP is that communication practices shape larger social structures. Therefore, CCP acknowledges the pipeline between education and the ballot box and the workforce. But if any given media reorganizes consciousness, relationships, and society, communication is inseparable from these processes. Therefore, communication and pedagogy are ongoing processes rather than endpoints.

The foundational understanding of CCP informs TPACK, which I take as an alternative to media literacy. CCP and TPACK allow me to integrate technologies into disciplinary practices and pedagogical commitments. Within the TPACK model, I focused on the technological pedagogical knowledge (TPK) component to align the availability and use of generative-AI tools with the course objectives (research writing).

Therefore, in addressing students’ knowledge deficit, CCP and TPACK centralize technology to analyze (discourse analysis; qualitative analysis) and enhance communication (creating relationships; structured analysis; storytelling). With CCP and TPACK, I focused on addressing students’ general knowledge about technologies, taking a logical, Socratic approach to thinking through problems.

Implementation

The project involves two parts. The first is data collection. For the classroom project, data was collected from customer reviews on Amazon’s website. Each student was responsible for collecting 5, 3, and 1-star reviews for coffee products; additional information for each product was also collected. These reviews were organized into an MS-Excel sheet.

The data collection process proved to be the most difficult because the average student is lazy and refuses to think. The students managed to collect data but there were far too many issues. I simply need to develop comprehensive instructions for this step in the process, too.

Following the data collection, students received step-by-step instructions on utilizing the MS-Excel sheet and ChatGPT. The instructions were listed in a PowerPoint presentation with screenshots. These instructions also listed the prompts for ChatGPT.

As the classroom project included three parts, the instructions covered one of those parts, providing students with the conditions and opportunities to experiment with ChatGPT on the other parts.

Assessment

The foremost form of assessment was delivered by students’ reactions to using ChatGPT with the instructions, exclaiming, “That’s what you’re supposed to do!” or “I’ve been using it wrong.” While I do not think there is a “correct” way to use ChatGPT, it is amusing to see students realize that there are, perhaps, “proper” and “useful” ways. Additionally, such reactions from frequent users of ChatGPT, such as those who use it at work, confirm the rampant media illiteracy in society.

I used both formative and summative assessments. The formative assessment was informal, during the instruction process in the classroom. These assessments were primarily troubleshooting in nature, allowing me to analyze students’ cognitive, affective, and psychomotor skills.

The summative assessment was formal, during the grading process. As part of the project, students submitted a worksheet of their chatlogs with ChatGPT and a “Method” section in their essay. The worksheet allowed me to review students’ prompts and steps on ChatGPT. However, even though students had the opportunity to experiment to an extent, no one deviated from the instructions to attempt something different. This is bad but, perhaps, also good.

The “Method” section of students’ essay was also part of the summative assessment. In the section, students detailed the steps to use ChatGPT, writing both macro- and micro-steps. Students effectively described their use of ChatGPT. Further, the subsequent “Analysis” section also demonstrated students’ ability to utilize ChatGPT for qualitative analysis.

The project provided value to me in two ways: First, language-modelling tools that can sieve large amounts of data ease out my work, especially concerning teaching qualitative analysis. Instead of delivering specific instructions to individual students, I can address the entire class to execute the same steps. From here, I can identify specific types of analyses that students want to conduct and organize resources (energy, time) accordingly.

Moreover, by conducting small scale projects with students, I can develop strategies for larger research projects. Thereby, linking research and teaching.

The project provided value to students, too, as two students used ChatGPT for their other projects in the same course. One student analyzed institutional health documents to identify course-related concepts. These documents were 10 pages long and would have been difficult for the student to manually analyze it within the deadline timeframe. Another student also analyzed institutional health documents, but worked with such a small dataset that using ChatGPT was not helpful.

Reflections and Next Steps

I want to conduct the project in the classroom again. I also want to use it in Senior Capstone to encourage students to do qualitative research with larger datasets. However, the dataset must be at least 10,000 words.

A similar project can be conducted focused on rhetorical fallacies. Students might use ChatGPT to identify fallacies in their own work. Thus, improving argumentation and writing.

What did I learn? Students do not know how to use basic tools, like ChatGPT, despite its mass availability and wide integration onto platforms, including cellphones. When students lack general skills, pedagogical tools, like Bloom’s Taxonomy and TPACK, are no longer helpful because students fall outside the available categories. As we pay attention to the development of generative-AI, we must also attend to our paradigms and theories of pedagogy.