Editorial Record: Submitted June 9, 2023. Revised September 20, and November 17, 2023. Accepted November 21, 2023. Published March 2024.

Author

ABSTRACT

With generative artificial intelligence – or generative AI – use on the rise, particularly in content creation, public relations students need to gain skills in applying such tools in PR writing. This article documents the author’s experience requiring the use of generative AI in writing assignments. The results, including research findings, are intended to help PR educators implement generative AI in their classrooms.

GIFT Overview

Rationale

Since the advent of generative AI systems, like the well-known ChatGPT (and GPT-3), both trade and academic articles have tried to make sense of its potential for and potential consequences in communication fields, like PR (Carufel, 2023; Luttrell et al., 2020). The author and two colleagues have also been researching the attitudes of PR educators, practitioners, and students about teaching emerging technologies (including generative AI). The preliminary findings of that research were a significant inspiration for the assignment detailed herein.

Additionally, a recent survey of executives found that 49% of their companies use ChatGPT, and 93% plan to expand its use (ResumeBuilder. com, 2023). There are, however, calls to pause its surge (Samuel, 2023).

Yet, Krishna et al. (2020) assert it is “imperative for [PR] educators to design courses and curricula to reflect the needs of the profession and prepare students … with the key tools and skills they need to be successful in their careers” (p. 4). And still, during a presentation about generative AI, an emerging technology expert said, “You won’t be replaced by AI. You’ll be replaced by someone who knows how to use AI” (A. Hood, personal communication, Nov. 22, 2022).

Connection to KSATs

Incorporating generative AI into our curricula is essential to support knowledge, skills, abilities, and traits (KSATs) identified in the Commission on Public Relations Education report by Duhe et al. (2018), namely writing, digital literacy and technology, and critical thinking.

Numerous studies, including Krishna et al. (2020), have confirmed writing is a vital PR skill. Furthermore, Floridi and Chiaratti (2020) concluded those who write professionally (e.g., PR practitioners) will need to use generative AI to prompt proper output, assemble, and edit it. Additionally, this author posits the ability to train AI models to produce writing in a particular voice is needed. Thus, if educators are to prepare students to be effective PR writers, they must ensure students can use generative AI. Xie et al. (2018) noted “a dire need to incorporate digital components in the curricula and believed it is important to transform students into digital thinkers” (p. 296). Indeed, Duhe et al. (2020) concluded digital technology is an area of “growing importance to the [PR] profession” (p. 63). Even more important than how to use generative AI, PR students must also develop the ability to discern when not to use generative AI. Indeed, Tóth et al. (2020) identified activities a digitally-competent person should be able to perform, including, “correctly applying digital technologies to solve various functional tasks; effectively using digital technologies and obtaining real practical results…self-confident[ly], critical[ly], and creative[ly] using digital technologies” (p. 153).

Assignment Learning Objectives

The generative AI assignments herein feature in an upper level PR writing course. The learning objectives (below) support, at higher levels of Fink’s taxonomy (2013), the overall course objectives:

• Recognize the strengths and weaknesses of generative AI.
• Create public relations writing using generative AI.
• Compare and contrast generative AI writing with human writing.
• Reflect on the use of generative AI in the writing process.

Connection to Public Relations Practice and/or Theory

Xie et al. (2018) found advertising and PR education was “… outdated in terms of preparing students for the changing industries, especially in terms of coping with emerging media technologies…” and respondents further believed digital skills “…should be embraced at every step in the program of study…” (pp. 302-303). Indeed, underscoring these points, the Commission for Public Relations Education recently highlighted the urgency of both addressing the lack of industry vision for AI and the need for entry-level graduates to possess AI skills and knowledge to prepare them for a field in which AI is being used in a variety of applications. (Toth & Bourland-Davis, 2023, pp. 27-28). The mandate for these skills will only grow as generative AI proliferates.

An analysis of PR job ads by Meganck et al. (2020), however, revealed “technology skills” were sought in just 4.1% of ads, while “social/digital media” appeared in 35.3% (p. 3). This is puzzling, given a comment by Panda et al. (2019) that “communication specialists are under pressure to unlearn old manual skills and learn new digital capabilities to survive in the new work environment” (p. 198).

The current environment feels like certain periods when PR had to adapt to “new digital capabilities” like email, websites, social media, and blogs. When those technologies debuted, PR practitioners evaluated benefits, costs, risks, and applications, preparing for when implementation became advisable or inevitable. The advent of generative AI is similar, so we must integrate it into the PR writing curriculum.

Class Testing

Following the imperative noted by Krishna et al. (2020) to prepare students for success (p. 4), the author also bore in mind the guidance of Galloway and Swiatek (2018), who wisely noted that PR practitioners need not be technology experts, but have a “…sufficient understanding of AI and related technologies…” so they can better advise clients (p. 736). The author required students to use Craftly, a generative AI tool for PR, advertising, and marketing writing, on four of five primary writing assignments (PWAs): backgrounders, blog posts and social posts, feature and social posts, and a letter to the editor. How students used Craftly varied by assignment (for example, see Figure 1). However, to establish a baseline of their writing ability, students wrote their first assignment – a news release – without AI.

As noted, each PWA has specific AI use requirements. Students were required to use Craftly as follows for each:

• Feature – outline the article, rewrite sentences, create quotes, or create a social media post to promote the article.
• Backgrounder – outline the document, rewrite sentences, or write the introduction.
• Letter to the Editor – outline the document, rewrite sentences, write the introduction, or create quotes.

In addition to the PWAs, students used Craftly for in-class learning activities to develop AI prompt-generating skills, including headlines, quotes, news release ledes, and similar output.

Evidence of Learning Outcomes

At the semester’s end, students completed a survey about their use of generative AI in class. The research plan and instrument received institutional review board approval. Of 10 students in the class, nine attempted the survey, and eight finished it. Due to the small class size, no demographic data were collected to ensure confidentiality. The author acknowledges the data set is small (as is often the case with scholarship of teaching and learning—or SoTL—research) and plans to continue this research in future course sections and grow the data set. There are, however, a few notable findings worth sharing that support this assignment.

Early-semester discussions about generative AI tools revealed students had little knowledge of generative AI’s uses, limitations, and risks. Therefore, the author wanted to know how student attitudes about AI changed as they used the technology, specifically its importance in their future careers, its danger, its ability to enhance PR practitioners’ effectiveness, ethical concerns about its use, and if understanding it would set them apart from other job candidates (see Table 1).

Table 1

Changes in Student Attitudes Regarding Generative AI/GPT-3

When considering if GPT-3 skills and their future careers the most remarkable increase was related to seeing the importance of GPT-3 to success in their future careers (t=-4.27, p=0.003) from the beginning of the semester (M=3.11, SD=1.17) to the end of the semester (M=4.56, SD=0.53). The findings suggest students see value in learning generative AI skills to help them advance in their careers and support the admonition of Krishna et al. (2020) to “prepare students to enter the workforce with the key tools and skills they need to be successful in their careers” (p. 4).

When asked to rate potential threats of GPT-3 (see Table 2; 1= is less of a threat to 5 = more of a threat), students expressed concern about loss of writing skills (M=3.75, SD=1.04), which suggests the need not only to teach the effective use of generative AI as a writing tool but also to ensure the ongoing development of fundamental writing skills. Further, with low consumer trust in AI-produced content (Owen, 2023), PR graduates knowledgeable in how to effectively employ generative AI will be valuable contributors to the new AI-enabled PR workforce.

Table 2

Student Attitudes About Generative AI/GPT-3 Use Threats in PR

A common misconception, however, is that younger generations are automatically more adept with and interested in new technologies. Indeed, Tóth et al. (2020) found while students possessed generally high levels of digital competence when switching to online learning during the COVID-19 pandemic, “… their knowledge and skills are limited to…basic office suite skills, emailing, using social sites, and surfing the Internet. They are thus digital users and consumers of content rather than its creators” (p. 152).

Given the sophisticated nature of generative AI technology, the author wanted to understand student attitudes regarding technology, generally, to understand how those attitudes might shape class teaching using generative AI. Research by Goode (2010a and 2010b) introduced the concept of a technology identity, which helps classify a person’s relationship with technology using an identity model to assess “four areas of an individual’s belief system, namely beliefs about one’s technology skills, opportunities, and constraints to use technology, the importance of technology, and one’s own motivation to learn more about technology” (2010a, p. 590).

Data collected (see Table 3) using Goode’s model (which the author adapted to use a 5-point Likert-type scale) revealed all students saw themselves as fluent in the use of technology and most were motivated to learn more about technology. This model can help educators understand potential barriers and opportunities before introducing generative AI technology in the classroom to use in an assignment such as this.

Table 3

Student Technology Identity Self-Report

Anecdotal student comments were, in many ways, more compelling and offered greater insight than the survey findings. This is an excerpt from one student’s introspection:

My final submission was a huge improvement from my draft and I really enjoyed this assignment. I also enjoy this course in general because it encourages us to utilize a brand-new tool that is frowned upon by others… Generally speaking, many teachers/professors were immediately against ChatGPT before even giving it a chance. I feel as though this is a tool that is supposed to benefit us and improve our writing skills.

Additional student comments included:

“I have a very difficult time editing my own work. It’s easier to edit [AI]. It feels less personal.”

“I want to start finding ways to implement AI into my life more. After using it in this class, I can see that it isn’t perfect, and it probably never will be, but there are some useful factors to it. I also do not see AI going anywhere anytime soon, so finding a way to work with it seems helpful for my future in PR.”

“I noticed that when I begin my papers, I tend to have trouble beginning the writing process. I found that Craftly can easily help me start and end ideas in a very descriptive but to-the-point way.”

“I also feel like I trusted Craftly a little bit too much. Although Craftly did help me develop my first [blog post] pretty much perfectly, it didn’t do the same for my second one. I should have only used Craftly for the brainstorming process…and crafted it myself. I think that the final results would have been much clearer and there would be less confusion using my own words and phrases.”

“The biggest thing for me…is the outlining. Sometimes I have a hard time getting started and organizing my thoughts. Craftly saves me at least 15 minutes.”

“I used to believe AI was just meant to be used for generating…silly responses, I hadn’t considered how useful it could be as a writing tool… AI has become one of my favorite ways of organizing my thoughts.”

… I had trouble making sure what the AI came up with was not an error. Craftly often sounded repetitive, and it made my writing a bit glib. In the beginning, it was harder for me to scrutinize and correct Craftly’s work, but by the end of the semester, I got much better at it.

…using AI is not something I thought I would ever be doing in a class. Before this class, I was very skeptical about AI as many others are. As someone with an interest in art and…the world of freelancing, the push for AI art almost put me off the idea completely. However, since using it in this class I am still skeptical but do not hate it as much as I did.

One student even expressed their feelings in poetry form:

I don’t trust AI, efficiency be damned.

Innocent at first but who knows where it’s going. Power beyond our control, some men have evil hands. Give the men time, we’ll reap what they’re sowing.

Minds will become grey, wastelands, reliant on something they could never be.

See, you and I don’t comprehend its power.

To destroy our lives in new and creative ways; a new tool for the bourgeoise.

Our cries go unheard as, our creativity, they devour.

Humans are unique, dreaming in a universe greater than our comprehension.

Innovation doesn’t end because we choose to reject a dreamless invention.

Important themes emerged in these and other comments, which included writing efficiency, learning how to work with generative AI, and distrust of/skepticism about generative AI.

The business case for generative AI becomes clear when considering the revelations about saving time: 15 minutes (and associated overhead costs) saved in the writing process is significant when multiplied across numerous projects and clients. As previously noted, students agreed faster content creation was an AI-use opportunity.

Students also noted AI can be repetitive and generate inappropriate tone. As the semester progressed, students became more adept at removing such flaws from their drafts.

The author believes healthy skepticism about generative AI is the most vital insight students gain by doing these assignments. The students’ comments expressing such skepticism reinforce the survey findings and recall Galloway and Swiatek’s (2018) caution, “the [PR] profession’s more immediate need to acquaint itself with the broader societal and economic AI impacts should be considered alongside the need to apply AI” (p. 736).

Template Assignment Guides

Blog Post Instructions

Draft two blog entries about topics of your choosing that affect or relate to your nonprofit. Consider:

• new laws or regulations
• social issues
• race issues
• demographic issues
• services offered
• challenges faces
• expertise to share Your blog posts will include:
• A related photo (with alt text)
• Meta description
• Word limit 800-1000 words

Writing Your Blog with Craftly

• Think of a brand subject your want your blog posts to be about (e.g. demographics)
  • Use Generate Blog Topics tool
• Select a topic for your first post, go to Blog Builder
  • Follow the steps and prompts to create the blog content
  • Edit and revise
• Select a topic for your second post, use the Blog Outline tool
  • Use the Blog Introduction to start your blog post
  • Use the other Craftly tools as much, or as little, as you wish
• Be sure to take notes about the process of using Craftly for your introspection and cite Craftly content in your blogs

For the images…

• Use Craftly to create at least one image. The other can be a stock image, screen grab, or original photo.
• Your alt text for each should be <125 characters and should describe what is seen in the image, as discussed in class. Remember the purpose of the alt text is both to describe the picture to someone using a screen reader and index the image for search engine retrieval.

For the social posts…

• Each blog post must include one social post to drive traffic to the blog:
  • You can choose Facebook, Instagram, Twitter, or LinkedIn, based on what is most relevant for your nonprofit.
  • Use Craftly to generate one of your social posts (identify which one).
  • You may use internet slang and/or emojis as appropriate for your nonprofit and the blog post messages.

Introspection Instructions

The introspection is a short piece that asks you to consider errors, word choice, clarity, adherence to the creative brief, generative AI use, and an overall reflection on your writing. Use the guide below to write the introspection. Because it is written, quality counts. See the attached rubric, too.

After each major writing assignment, you must complete an introspective reflection on the assignment from draft to final. Step away from the assignment for a day or so. Then re-read it and respond to the following:

Errors

• Were there any surprise errors? What about consistent errors?
• Are there words you regularly misspell? What are they? How can you avoid in the future?

Word Choice

• Were any word choices called out in the feedback? How would you resolve them?
• Looking at your final with fresh eyes, are there any other word choices you would change?

Clarity

• Were any sentences called out in feedback for lack of clarity? How would you resolve them?

Creative Brief

• Did you miss any requirements specified in the creative brief? Why do you think that happened?

Overall

• Upon re-reading, did you find other improvements you could make?
• How did you feel about your draft compared with your final?
• Was there instructor feedback you found valuable? Confusing?
• What were your thoughts on using AI for the assignment?
• Were there any challenges you faced when writing the assignment? Technical process, inspiration, other.
• How will you resolve those challenges for the next assignment?

Assignment Grading Criteria or Rubric

This course uses ungrading principles, relying heavily on feedback (Blum, 2020). Students submit each PWA first draft for instructor feedback. Students do not receive a letter grade for each PWA. Instead, they work to improve their writing through feedback until it is of A quality (16-18 points) based on a rubric (see Figure 1).

Figure 1

Primary Writing Assignment Rubric

Each student then works with a pre-assigned peer review partner to critique one another’s work. Students then submit a final draft (subject to further revision) noting any AI content (see Figure 2).

Figure 2

Example of Student Canvas Submission Identifying AI Use

The introspection rubric (see Figure 3) assesses students’ ability to interrogate and better understand their writing process, as well as examine their own use of generative AI.

Figure 3

Introspection Rubric

Teaching Note

These assignments are appropriate in upper-level PR writing courses because they are as much about how to not use generative AI as they are about how to use it. Students should have a foundation in PR writing, grammar, AP style; law and ethics; and PR theory.

Generative AI use varied by assignment and was explicitly defined.

The instructions for a pair of blog posts and an associated social media post are in the Template Assignment Guide. Introspections were also required because they “…have the metacognitive function of helping writers become more aware of their own thinking processes” (Bean & Melzer, 2021, p. 88). Such awareness is essential when writing with generative AI, as illustrated by the student comments.

Based on the author’s experience with these assignments in the classroom, instructors should consider the following when deciding how or if to implement generative AI use in their classrooms:

Generative AI Tool Knowledge

• Become power users of the chosen generative AI tool. Instructors should understand how to navigate the tool, know its strengths and weaknesses, and be able to provide instruction in its use and troubleshoot student problems.
• Train students thoroughly in how to use the chosen AI tool. The author’s students received training from Craftly.
• Consider assessing student technology identities using Goode’s model to understand the students’ orientation and motivation regarding tech use. It can also help to identify peer technology champions who could help their fellow students.

Generative AI Use

• Prescribe a variety of specific ways to use AI. As noted, students were required to use Craftly to outline, generate quotes, create social media posts, write blogs, craft headlines, and even create images. Allow them to create no- or low-stakes content before using the tool for a graded assignment. Also, require students to reflect on their use of AI and assess AI’s strengths and weaknesses.
• Require students to identify AI output used (see Introspection Instructions) and caution students against using AI in other classes, as it may be prohibited.

Course Content

• Teach best practices for completing each writing assignment, irrespective of AI use.
• Review effective proofreading and editing techniques, focusing on identifying AI-generated shortcomings. Teach shortcomings of AI writing (e.g., redundancy, factual errors, lack of AP style use, and inconsistent or inappropriate tone) to help them effectively edit AI output. Finally, pair each student with a peer editor to support the revision process.

Any generative AI tool should work for these assignments. A paid platform is ideal, as free or “freemium” versions may have word limits or restrictions that may hinder students’ ability to complete assignments and affect the instructor’s assignment parameters. Craftly proved to be an effective tool.

Institutional IT policies may stipulate accessibility requirements for any AI tool chosen (or even prohibit their use). The author’s university requires all institution-purchased software to meet Americans with Disabilities Act (ADA) standards. Smaller software companies, especially start-ups, may not prioritize such compliance in early go-to-market plans. Instructors should verify IT requirements before comparing platforms and understand how long university approval and testing processes take to avoid implementation delays.

Opposition from colleagues or the administration is certainly possible. The author’s department supported experimentation with generative AI, but not all institutions will have the same attitudes.

While not a guarantee of acceptance, crafting what Fink (2013) refers to as “significant learning experiences” (p. 7) that comprise (a) meaningful learning objectives, (b) teaching and learning activities, and (c) feedback and assessment, as well as applying appropriate academic rigor are, in the author’s opinion, essential precursors to reassure skeptics. Another important reminder: Generative AI is just a tool. Knowing how to use it does not make one a writer any more than knowing how to use Adobe Illustrator® makes one an artist.

Conclusion

Incorporating mandatory generative AI use in a PR writing course initially seemed risky. However, it also seemed risky not to incorporate the technology given the incredible attention ChatGPT and other generative AI tools were receiving in the media.

Students seemed surprised to be trusted with using generative AI on school writing assignments. It was as if they were being offered a secret code granting them special powers. The author thinks this is true and believes they will be among the first to enter the PR workforce having had guidance in how to (and how not to) use generative AI in PR writing.

References

Bean, J. C., & Melzer, D. (2021). Engaging ideas: The professor’s guide to integrating writing, critical thinking, and active learning in the classroom (2nd ed.). Jossey-Bass.

Blum, S. D. (Ed.). (2020). Ungrading: Why rating students undermines learning (and what to do instead) (1st ed.). West Virginia University Press.

Carufel, R. (2023, April 26). Will AI ultimately replace content writers? Nearly 7 in 10 content marketers believe it will over the next five years. Agility PR. https://www.agilitypr.com/pr-news/public-relations/will-ai- ultimately-replace-content-writers-nearly-7-in-10-content- marketers-believe-it-will-over-the-next-five-years/

Duhe, S., Ferguson, D., Halff, G., & Shen, H. (2017). Undergraduate Curriculum: Courses and Content to Prepare the Next Generation of Public Relations Practitioners. In E. L. Toth & K. Lewton (Eds.), Fast Forward: The 2017 Report on Undergraduate Public Relations Education (pp. 59-63). Commission on Public Relations Education. http://www.commissionpred.org/wp-content/uploads/2018/04/ report6-full.pdf

Fink, L. D. (2013). Creating significant learning experiences: An integrated approach to designing college courses. John Wiley & Sons, Inc.

Floridi, L., & Chiriatti, M. (2020). GPT-3: Its nature, scope, limits, and consequences. Minds and Machines, 30(4), 681–694. https://doi.org/10.1007/s11023-020-09548-1

Galloway, C., & Swiatek, L. (2018). Public relations and artificial intelligence: It’s not (just) about robots. Public Relations Review, 44(5), 734–740. https://doi.org/10.1016/j.pubrev.2018.10.008

Goode, J. (2010a). Mind the gap: The digital dimension of college access. The Journal of Higher Education, 81(5), 583–618. https://doi.org/10.1353/jhe.2010.0005

Goode, J. (2010b). The digital identity divide: How technology knowledge impacts college students. New Media & Society, 12(3), p.497–513. https://doi.org/10.1177/1461444809343560

Krishna, A., Wright, D., & Kotcher, R. (2020). Curriculum rebuilding in public relations: Understanding what early career, mid- career, and senior PR/communications professionals expect from PR graduates. Journal of Public Relations Education, 6(1), 33-57. https://aejmc.us/jpre/wp-content/uploads/sites/25/2020/01/ KRISHNA-ET-AL-JPRE-6.1-Curriculum-Rebuilding-in-PR.pdf

Luttrell, R., Wallace, A., McCollough, C., & Lee, J. (2020). The digital divide: Addressing artificial intelligence in communication education. Journalism & Mass Communication Educator, 75(4), 470–482. https://doi.org/10.1177/1077695820925286

Meganck, S., Smith, J., & Guidry, J. P. D. (2020). The skills required for entry-level public relations: An analysis of skills required in 1,000 PR job ads. Public Relations Review, 46(5), 101973. https://doi.org/10.1016/j.pubrev.2020.101973

Owen, J. (2023, June 1). Fear and confusion over brands using AI, report finds. PRWeek. https://www.prweek.com/article/1824756/fear-confusion- brands-using-ai-report-finds

Panda, G., Upadhyay, A. K., & Khandelwal, K. (2019). Artificial intelligence: A strategic disruption in public relations. Journal of Creative Communications, 14(3), 196–213. https://doi.org/10.1177/0973258619866585

ResumeBuilder.com. (2023, February 27). 1 in 4 companies have already replaced workers with ChatGPT. https://www.resumebuilder.com/1-in-4-companies-have-already- replaced-workers-with-chatgpt/?utm_source=newsletter&utm_ medium=email&utm_campaign=newsletter_ axioscommunicators&stream=business

Samuel, S. (2023, March 20). The case for slowing down AI. Vox. https://www.vox.com/the-highlight/23621198/artificial- intelligence-chatgpt-openai-existential-risk-china-ai-safety- technology

Toth, E. L., & Bourland-Davis, P. (Eds.). (2023). Navigating Change: Recommendations for Advancing Undergraduate Public Relations Education. Commission on Public Relations Education. https://www.commissionpred.org/wp-content/uploads/2023/10/ CPRE-50th-Anniversary-Report-FInal-Final.pdf

Tóth, T., Virágh, R., Hallová, M., Stuchlý, P., & Hennyeyová, K. (2022). Digital competence of digital native students as prerequisite for digital transformation of education. International Journal of Emerging Technologies in Learning, 17(16), 150–166. https://doi.org/10.3991/ijet.v17i16.31791

Xie, Q., Schauster, E., & Neill, M. S. (2018). Expectations for advertising and public relations education from agency executives: A comparative study between China and the United States. Journal of Current Issues & Research in Advertising, 39(3), 289–307. https://doi.org/10.1080/10641734.2018.1490358

© Copyright 2024 AEJMC Public Relations Division

To cite this article: Ryan, S. (2024). Generative AI in the classroom: Teaching a critical skill for the public relations writing student. Journal of Public Relations Education, 9(3), 25-45. https://journalofpreducation.com/?p=4241