Laura DeLuca

Graduate Student
Dietrich College of Humanities and Social Sciences
Spring 2025

76-101 Cleopatra’s Cultural Afterlife: Representations of Ancient Egypt in Popular Discourse (14-week course)

Research Question(s): 

1. To what extent does engaging with a genAI activity on ancient Egyptian culture change students’ attitudes toward cultural biases, historiography, and the humanities?
2. To what extent does students’ self-efficacy for critically examining media change after engaging with the genAI activity?

Teaching Intervention with Generative AI (genAI):

DeLuca integrated genAI into a low-stakes, in-class activity in which students prompted genAI (Microsoft Copilot) to generate text- and image-based representations of Cleopatra and Ancient Egypt. Over the course of one class session, students created representations of cultural aspects that interested them, such as fashion, daily life, religious practices, etc. They then answered a series of questions about the genAI’s output that required them to reflect on the accuracy of the representations, what stereotypes emerged, whether such stereotypes were present in other course material, and what misconceptions could arise from these stereotypical representations. Prior to completing the genAI activity, DeLuca provided a lesson on prompt engineering.

Study Design:

DeLuca had all students engage in the genAI activity (there was no comparison group). Students completed a reflection survey immediately before and a week after the activity, which measured their attitudes and their self-efficacy for critically evaluating media and cultural portrayals.

Sample size: 18 Students

Data Sources:

1. Pre/Post survey about student attitudes about cultural biases, historiography, and the humanities, as well as self-efficacy for critically examining media

Findings:

1. RQ1: None of the attitude measures significantly changed after engaging with the genAI activity, nor did the aggregate attitude.
   
   

   Figure 1. Mean attitudes were the aggregate of attitudes toward cultural bias, historiography, and the humanities. There was no significant change in attitudes from pre to post after engaging with the genAI activity, t(17) = .18, p = .86. Error bars are 95% confidence intervals for the means.

2. RQ2: Overall self-efficacy for critically analyzing media did not significantly change from pre to post.
   
   
3. 
   

   Figure 2. Mean self-efficacy was the average of responses across the five self-efficacy items. There was no significant change in self-efficacy from pre to post after engaging with the genAI activity, t(17) = -1.02, p = .32. Error bars are 95% confidence intervals for the means.

Eberly Center’s Takeaways:

1. RQ1&2: GenAI is known to perpetuate stereotypes and biases in its outputs. Stereotypical text- and image-based depictions of ancient Egypt by Copilot were verified by instructor testing and artefacts submitted by students following the activity. Over the course of a week, students did not change in their attitudes about cultural representations, the subjectivity of history, or the value of the humanities after completing an activity that required them to create cultural representations of Ancient Egypt using genAI. Despite this, students entered the activity scoring high on these beliefs, consistent with the learning goals for the activity. Similarly, students’ self-efficacy for critically examining media was high at the outset and did not change after the activity. It is possible that students need further scaffolded exposure to potentially flawed media representations and opportunities to critique them to further develop these attitudinal beliefs and confidence in their skills. It is also possible that earlier course experiences had already shifted these beliefs to the measured high levels. One benefit of genAI in this study was that it allowed students to personalize their activity experience by exploring an aspect of ancient Egyptian culture that they were most interested in. How this personalization impacted students remains an open question.