Abstract
Generative AI may significantly disrupt the teaching and practice of business communication. This study of 343 communication instructors revealed a collective view that AI-assisted writing will be widely adopted in the workplace and will require significant changes to instruction. Key perceived challenges include less critical thinking and authenticity in writing. Key perceived benefits include more efficiency and better idea generation in writing. Students will need to develop AI literacy—composed of application, authenticity, accountability, and agency—to succeed in the workplace. Recommendations are provided for instructors and administrators to ensure the benefits of AI-assisted writing can outweigh the challenges.
Introduction
Many scholars and experts have suggested for years that AI is poised to significantly disrupt education and work in the future (Getchell et al., 2022). Yet, the public release of ChatGPT by OpenAI in November 2022 has led to many scholars and experts to conclude that AI-assisted writing is a reality for the here and now. Various scholars and experts refer to ChatGPT as a “game-changer” (van Dis et al., 2023, p. 224), a “tipping point” (Mollick, 2022), and “transformational” in the evolution of AI tools. Bill Gates referred to the transformational nature of ChatGPT as “every bit as important as the PC, as the internet” (Nerozzi, 2023). Microsoft CEO Satya Nadella suggests it will dramatically increase productivity for the world’s 8 billion people (Schechner, 2023). Google CEO Sundar Pichai, in an announcement of Bard, stated that AI will “significantly improve billions of lives” and that Bard will be an “outlet for creativity” and a “launchpad for curiosity” (Pichai, 2023). Meta CEO Mark Zuckerberg also announced a competing product called LLaMA (Leswing, 2023). Most recently, Elon Musk, Tesla CEO and Twitter owner, registered a firm named X.AI Corp, an artificial intelligence start-up that will rival ChatGPT-maker OpenAI (Jin, 2023).
While many tech corporate leaders promote the transformational and positive nature of these tools, many experts and thought leaders worry about the potential negative impacts of AI. Some experts predict major job losses and job displacements (Lowrey, 2023), and some survey work suggests many companies are already replacing people with ChatGPT (Williams, 2023). Kissinger et al. (2023) worry that it can worsen international relations if some nations monopolize these technologies and in effect enact a new form of imperialism. Moreover, regulators have been calling for rules ahead of mass adoption of ChatGPT. EU’s privacy watchdog, European Data Protection Board (EDPB), has a task force in place to develop common policy for AI. Italy was the first EU country to ban ChatGPT until regulation is in place, and Germany and Spain are investigating potential data breaches of ChatGPT (Sterling, 2023).
Some experts forecast that business communication will be dramatically influenced by AI-assisted writing and other forms of generative AI. Davenport and Mittal (2022) predict the following: We are now only scratching the surface of what generative AI can do for organizations and the people within them. It may soon be standard practice, for example, for such systems to craft most or all of our written or image-based content—to provide first drafts of emails, letters, articles, computer programs, reports, blog posts, presentations, videos, and so forth.
If forecasts such as this prove true, members of the business communication community will need to reconceptualize how to teach many aspects of business communication. To this point, many scholars, experts, and journalists are editorializing about the potential influence of such tools. Many of these early analyses of ChatGPT involve demos of what the tool can and cannot do. Yet, little is known about how communication instructors are reacting to ChatGPT. This study of business communication instructors is the first known and large-scale exploration of how higher education instructors view the emergence of these tools. Specifically, the purpose of our study is to highlight the current concerns and anticipated benefits of AI-assisted writing shared by business communication instructors. The results of this study suggest potential approaches to capturing the benefits of generative AI while addressing the substantive concerns. Broadly, the purpose of this study is to capture the collective wisdom of business communication instructors to develop an expert-driven framework of AI literacy for communicators. This model of AI literacy can help communicators in interpersonal, team, and leadership contexts.
Literature Review
A small set of scholarly work has focused on the role AI plays in business communication (e.g., Brown-Devlin et al., 2022; Cardon et al., 2021; Getchell et al., 2022; Men et al., 2022). No known research in business communication has addressed generative AI (including ChatGPT) in depth. This is not surprising given the quite recent public release of ChatGPT and its dramatically superior abilities compared to prior chatbots. As a foundation for our study, we briefly cover generative AI and AI-assisted writing as well as generative AI and the academic environment.
Generative AI Adoption and AI-Assisted Writing in the Workplace
Scholars and experts classify AI technologies in many ways (Getchell et al., 2022). ChatGPT is an AI-powered chatbot that can be classified as generative AI and is built on a large language model (LLM). van Dis et al. (2023) explain that “ChatGPT has caused excitement and controversy because it is one of the first models that can convincingly converse with its users in English and other languages on a wide range of topics” (p. 224). Compared to other AI systems, it is significantly better at understanding questions from people, providing accurate responses, and responding in human-like language. It has many potential applications in the workplace, including in marketing and sales, customer service, operations, information technology, engineering, legal analysis, finance, and R&D (Chui et al., 2022; Davenport & Mittal, 2022). We classify ChatGPT as AI-assisted writing for the purposes of this research project because of our focus on text output and writing courses. Mollick (2022) suggests the transformative potential of ChatGPT and similar tools is that they can complete creative tasks with little risk of costly failure: Until now, AI has primarily been aimed at problems where failure is expensive, not at tasks where occasional failure is cheap and acceptable—or even ones in which experts can easily separate failed cases from successful ones. A car that occasionally gets into accidents is intolerable. But an AI artist that draws some great pictures, but also some bad ones, is perfectly acceptable. Applying AI to the creative and expressive tasks (writing marketing copy) rather than dangerous and repetitive ones (driving a forklift) opens a new world of applications.
Many writing instructors view writing as a representation of effective thought. Generative AI appears to possess imperfect but relatively strong reasoning skills. Bang et al. (2023) explored the ability of ChatGPT to engage in 10 different reasoning categories. They found that it was about 64 percent accurate in these categories, performing best in logical (i.e., deductive, inductive, abductive) reasoning and less effectively at nontextual semantic (i.e., temporal, spatial, mathematical) reasoning and commonsense reasoning. They also showed that it engages in intrinsic and extrinsic hallucination. They concluded, “ChatGPT exhibits a tendency to be a lazy reasoner and that its capabilities are inconsistent across various reasoning abilities.” They conclude that responsible generative AI must address issues of fairness and demographic bias.
One indicator of the potential value of ChatGPT’s capabilities is the adoption rate. ChatGPT reached 100 million users in just two months, surpassing TikTok and Instagram as the fastest growing web platform in history (Chow, 2023). Assessing how much generative AI (including ChatGPT) is being used in everyday communication in the workplace is difficult to assess. Emerging survey work and anecdotal accounts suggest it is common. A survey of 621 business leaders in the United States who are early adopters of the technology in February 2023 reported that ChatGPT is being used for writing job descriptions in 77% of companies; drafting interview questions, 66%; responding to job applicants, 65%; copywriting/content creation, 58%; customer support, 57% of companies; summarizing meetings and corporate documents, 52%; research, 45%; and generating task lists, 45% (Williams, 2023). A survey study of working adults in the United States found that roughly 42% were using ChatGPT to research a topic or generate ideas for work, 32% for drafting messages, 26% for drafting reports, and 22% for editing text (Cardon et al., 2023).
ChatGPT has many limitations. While ChatGPT relies on an enormous and growing data set of human language, it currently does not always have access to the most up-to-date information, may generate inaccurate information, and may be biased (Deng & Lin, 2023). For example, Hartmann et al. (2023) showed that ChatGPT is biased for pro-environmental, left-libertarian political orientations. Concerns related to ChatGPT and other conversational AI include plagiarism, security, and fake news (Guo et al., 2023).
Many scholars have discussed the role of AI-assisted writing in science. van Dis et al. (2023) believe that ChatGPT will become integrated into all aspects of research writing: summarizing research literature, drafting and improving research manuscripts, conducting statistical analysis and detecting problems with analyses, designing experiments, writing entire manuscripts, and even be integrating into the editorial process with accept and reject decisions. They foresee that generative AI tools like ChatGPT could make research faster and more efficient, more innovative, more equitable, and more diverse. Yet, it could also lead to lower transparency, less autonomy for human researchers, bias, and the spread of misinformation. They recommend five steps to using generative AI: (1) hold on to human verification; (2) develop rules for accountability; (3) invest in truly open LLMs; (4) embrace the benefits of AI; and (5) widen the debate. Other scholars suggest a similar set of ethical guidelines (Liebrenz et al., 2023).
Emerging research suggests that for many types of tasks, human-AI collaboration can outperform human-only or AI-only decision making (Fügener et al., 2021). Yet, recent research also shows that in some human-AI environments, group performance and human individuality may suffer because of the loss of unique human knowledge. Even though individuals may perform better with the assistance of AI in the short term, reliance on AI by individuals leads to groups of people without AI performing better than groups of people with AI in the long term because people lose complementary skills that support higher human-AI performance (Fügener et al., 2022). As a result, leaders and manager should understand principles for effective human-AI interaction and consciously explore potential roles of AI. Grounding their work on Siddike et al. (2018) and Babic et al. (2020), Getchell et al. (2022) built a framework of potential AI roles that could define the nature of AI-human collaboration. The role of AI in business communication can be as a tool, an assistant, a monitor, a coach, or a teammate.
Generative AI and the Academic Environment
There has been prominent coverage about ChatGPT in the context of higher education (D’Agostino, 2023; Rudolph et al., 2023). Most scholarly work on ChatGPT has been on academic integrity (Cotton et al., 2023). Many of these studies suggest evaluation will become particularly difficult given the challenge of discerning student work from AI-generated content (Swiecki et al., 2022).
Many scholars have already published demos of good output from ChatGPT for academic environments. Gilson et al. (2022) showed that ChatGPT functions fairly well on medical licensing exams and achieves the equivalent of a typical third-year medical student. An operations professor at Wharton used ChatGPT to demonstrate how it would do on a final exam for an MBA course at one of the top business schools in the world. He concluded that ChatGPT, in its current form, would receive a B or B– on his exam (Terwiesch, 2023). For scientific articles, academic reviewers only identified 63% of fake abstracts submitted by ChatGPT (Thorp, 2023).
Qadir (2022) demonstrated that ChatGPT can effectively produce high-quality, university work for tasks such as code generation, responding to conceptual questions, answering mathematical questions, creating essay outlines, developing stories with imaginative dialogues, and even producing effective persuasive writing. He concludes: Paradoxically, with advancing technology, classical human skills and liberal arts such as critical thinking, communication, and problem-solving become more rather than less important. These skills are essential for being able to effectively use and analyze information and technology, as well as for creating original and innovative solutions to complex problems. In today’s world, it is more important than ever to be able to think critically and analyze arguments, spot errors and misinformation, and make fixes when necessary.
Susnjak (2022) examined the use of ChatGPT for a variety of essay-type questions for online exams. He used ChatGPT to accomplish the full cycle of creating and evaluating high-order, open-ended exam question: (1) generate questions that require critical thinking; (2) respond to these questions; and (3) critically evaluate the responses to these questions. He followed this process in the following disciplines: sciences, education, humanities, and business. The output was evaluated in terms of Paul’s (2005) model of critical thinking (with the addition of persuasiveness and originality added): relevance, clarity, accuracy, precision, depth, breadth, logic, persuasiveness, and originality.
Yet, no known research has broadly explored how communication educators view the challenges and benefits of generative AI. Many opinion pieces by individual instructors exist with demos of various uses and misuses of ChatGPT. The value of a broad study of communication instructors is that it can tap into the collective wisdom and sentiments of these experts on business communication. These insights from communication experts can support a framework of artificial intelligence literacy for communicators in interpersonal, team, and leadership contexts.
Methodology
Since generative AI tools such as ChatGPT have not been widely used among publics or in the workplace until the last few months, this is considered an exploratory study. As experts in communication in applied settings, business communication instructors are an important group to understand the potential implications of how generative AI may influence communication instruction. We posed the following research questions:
What concerns do business communication and writing professors hold about AI-assisted writing (e.g., Chat GPT)?
To what degree do business communication instructors experience anxiety towards teaching or integrating AI-assisted writing into their courses?
What perceived benefits do business communication and writing professors hold about AI-assisted writing (e.g., Chat GPT)?
To what degree do business communication professors perceive AI-assisted writing (e.g., ChatGPT) will be accepted in the near future for workplace writing?
We developed a survey to address these research questions (see Appendix A). The survey was a time-consuming survey that required respondents to share their views in roughly 10 to 15 minutes. Because not all business communication instructors have personal experience with ChatGPT, the first 3 to 5 minutes of the survey allowed survey respondents to see ChatGPT output to three forms of prompts:
“What are the best ways to deliver bad news to employees?” (a simple query with advice)
“Please write a message to employees about a new policy that requires them to return to the offices for work.” (a workplace message)
“Write a 5 paragraph essay about the importance of addressing mental health in the workplace. Provide citations.” (a short essay with sources)
Next, survey respondents answered closed questions about concerns and perceived benefits of AI-assisted writing. These items were grounded in commentary and analysis of GPT from a variety of popular press articles. To address how business communication instructors view potential use of AI-assisted writing in the workplace, we adopted a technology acceptance framework. For our measures of technology acceptance, we adapted items from Venkatesh et al.’s (2003, 2012) UTAUT models. We also used Venkatesh et al.’s items for anxiety measures. The UTAUT constructs we addressed were performance expectancy (PE) or usefulness, effort expectancy (EE) or ease of use, social influence (SI), behavioral intention (BI), and anxiety (ANX). These items were highly reliable, with Cronbach’s alphas for this study at the following: PE, .91; EE, .82; SI, .87; BI, .96; and ANX, .79. Finally, because of the exploratory nature of our study, we had three open-ended items for business communication instructors to comment on perceived challenges and opportunities of AI-assisted writing.
We sampled from a large group of business communication instructors. Business communication instructors are an ideal group of instructors to survey because of their expertise in aligning teaching with contemporary business practices (Clokie & Fourie, 2016; Coffelt et al., 2022). Over the years, we have compiled a list of 1,800 business communication instructors. Altogether, 343 instructors completed the survey. Roughly 84% of these instructors are currently teaching business writing, 70% are teaching business presentations, 58% are teaching business reports, and 33% are teaching other writing topics. About 81% of the sample is from the United States, with the remainder coming primarily from Canada, Europe, and Asia. Roughly 61% of respondents reside in a business school, 21% in liberal arts/English, 10% in Communication, and 9% percent in other schools. Roughly 89% teach undergraduate students, and 41% teach graduate students. About 5% have taught from 1 to 4 years, 14% have taught 5 to 9 years, 39% have taught 10 to 19 years, and 42% have taught over 20 or more years.
Although ChatGPT is a new technology, instructors in our sample appear to have learned about it in many ways. About 83% have read articles about it; 75% have talked with colleagues about it; 46% have watched news about it; 37% have used it; 23% have watched online demos (e.g., YouTube videos) of it; 18% have participated in committees or other formal work groups to discuss how to use it in the classroom; 9% have attended workshops about it; and 4% have created policy about it. When asked, Overall, how much do you think you know about ChatGPT compared to other professors in your area? (on a scale from 1 = nothing to 5 = a lot), the average across the group was 3.0.
Participants responded to the following open-ended items: (a) What do you view as the primary challenges posed by ChatGPT for writing instruction and learning? How do you expect to address these challenges? and (b) What do you view as the opportunities associated with ChatGPT for writing instruction and learning? The open-ended comments were coded in three rounds. In the first round, each member of the research team independently read the comments and developed tentative ideas for codes and subcodes. The team then met and discussed a common approach to coding and developing definitions for each of 21 codes (see Tables 3 and 6). In the second round of coding, four members of the research team coded the comments. Each comment was assigned to two members of the research team in various configurations. Interrater reliability for the individual codes ranged from 77% to 100% on individual codes, with an overall average interrater reliability of 89%. In the third round, pairs of team members assigned to each comment discussed and reconciled any codes they had interpreted differently in Round 2.
We wanted the open-ended items to provide more richness to the survey findings. We specifically asked respondents to answer the questions briefly for two reasons: (1) to encourage higher participation and (2) to draw out what they viewed as the most salient issues. Of the 343 participants, 299 responded to the open-ended items. The average response was 5 sentences and 76 words long. Ultimately, after three rounds of coding, the codes were refined and interpreted in the context of larger categories and themes (Creswell & Creswell, 2018; DeCuir-Gunby et al., 2011; Merriam & Tisdell, 2016). Table 1 displays the codes and counts of codes. Additional details and examples of coded statements are available in Table B1 in Appendix B. Throughout the remainder of the article, we use representative quotes from participants. Following each quote, we refer to participants in the order they completed the survey (e.g., Participant 34 is referred to as P34).
Codes and Themes in Qualitative Analysis.
Note. Percentage is based on number of respondents who mentioned this code among the 299 respondents who completed the open-ended items.
A major goal of this study, particularly from the open-ended items, was to develop a framework of AI literacy that involves guiding questions or heuristics for communicators (van der Geest & Spyridakis, 2000; Getchell et al., 2022). A variety of frameworks of AI literacy exist; however these frameworks primarily focus on the technical aspects of AI use (Ng et al., 2021). One unique aspect of generative AI tools is use in everyday business communication tasks, which are primarily social in nature (Cardon et al., 2023). In their review of work about AI and business communication, Getchell et al. (2022) specifically call for scholars to develop frameworks to discuss
AI-assisted communication. Beyond technical and practical aspects of AI literacy, they call for frameworks that explore the ethical and social implications of AI-assisted business communication. Public discourse on generative AI has thus far been heavily influenced by software vendors who emphasize the technical aspects and productivity outcomes of generative AI or leaders or public figures who focus on the societal-level ethical implications. By relying on experts in business communication, we build a framework of AI literacy that focuses more so on how communicators seek to interact in interpersonal, team, and leadership contexts.
We acknowledge our positionality in this project (Creswell & Creswell, 2018). Generally, we view generative AI as technology that will impact how nearly all professionals communicate. We do not hold an overarching normative judgment about generative AI—we simply believe it is inevitable that generative AI will become integrated into the everyday communication activities of nearly all business professionals. Each of us is awed by the power and potential of the technology to enhance work productivity and creativity while also recognizing significant challenges the technology poses to human relationships and human development. Each member of this research group has attempted for many years to experiment with and research the impact of new technologies on business communication. We are likely considered early adopters by others. We would refer to ourselves as technology pragmatists who seek to understand the knowledge, skills, and attitudes necessary to be included in business communication curricula that will prepare students to become successful professionals. We also hold an underlying trust in the community of business communication scholars and instructors to produce collective wisdom about how to approach new forms of technology-mediated communication while honoring timeless principles of business communication.
Findings
The goals of our study were to broadly explore business communication instructors’ perceptions of AI-assisted writing in terms of challenges and opportunities. First, we present one of the most common views of business communication instructors: business communication instruction must change. Then, we organize our findings into the following sections: concerns with AI-assisted writing, policy making and equity considerations for AI-assisted writing, and perspectives about future use of AI-assisted writing in the workplace. The qualitative and quantitative analysis highlighted key abilities in the context of AI-assisted writing: application, authenticity, accountability, and agency. Application involves an understanding of AI tools and how well they align with communication tasks. Authenticity involves focus on genuine communication and prioritizing the human element. Accountability involves taking responsibility for the accuracy and appropriateness of AI-generated content and using generative AI in a fair and equitable manner. Agency involves professionals retaining control to make their own choices. We refer to these abilities throughout the Findings section because they represent the collective views of business communication instructors as far as key abilities for the AI Age, and we develop them into a framework for teaching about AI-assisted writing that we refer to as AI literacy. We describe this framework for AI literacy in the Discussion and Recommendations section.
Business Communication Instructors Must Change Their Teaching
Among the most salient observations from instructors is that they needed to change their teaching to integrate AI-related content into their courses. Instructors often reasoned that these tools should be addressed because they will be widely used in the workplace and that students are using them outside the classroom. For example, an instructor explained: Faculty should be thinking about how to effectively leverage these tools rather than obsessing about how to avoid/ban/regulate their use (because that is a losing battle). Also, if students use these tools in the classroom, they will almost certainly use them outside the classroom and possibly at work. How can we help emerging business professionals use tools like these effectively and ethically? (P86)
Instructors largely accepted the reality that AI should be integrated into communication coursework, yet they varied substantially in their sentiment about the change. Some viewed the change with enthusiasm. For example, an instructor stated: Rethinking priorities in the course as a whole to take advantage of this new tool! It’s a bit of a fashion statement now, but I think it can be, if not a game-changer in our field, at least a nudge to greater creativity in those preparing students to thrive in the workplace. (P105)
Others did not reveal their sentiments about the change, but simply stated the reality in matter-of-fact terms, often resolving to make the best of the change. For example, an instructor commented, “Rather than pretend it doesn’t exist, we need to learn about it and teach students how to use it so they can use it for good, not evil” (P71). Another characterized the change this way: “We need to face that automation has arrived and will be better and more powerful. We shall use it openly, learn and teach about it, and establish rules and ethics for its use” (P106). Others emphasized ChatGPT represented similar revolutions in the past. For example, an instructor said, “We have adapted—going from an abacus to a calculator, an encyclopedia to Google. We will continue to accelerate through technology disruption” (P239). Instructors emphasized that the attitude they adopted in this transition was important. For example, an instructor explained, “Instructors should not fear it. It can be the starting point for creative writing, just like Wikipedia is the starting point for deeper research on many topics, academic or otherwise” (P242).
Instructors provided several reasons for resistance to this change. Some resist because they worry about their ability to adapt based on their existing knowledge or skill sets, or simply because they don’t have time to get up to date. For example, an instructor commented, “I feel that technology is advancing at warp speed, and I worry I don’t have the time to create new assignments or discussions that can show students how to use it in their professional lives as a tool but not as a replacement for communication skills development” (P141). Another instructor explained, “It is changing and developing faster than I can develop lessons about it” (P318). As a result, instructors described a need for support, particularly the application or technical aspects of generative AI. For example, an instructor explained, “There will need to be more training, especially for teachers, to have a better understanding of this tool and how it benefits the student” (P285).
Many instructors resist the change because of what they view as problems without clear solutions. The most common of these problems involves distinguishing student work from AI-generated work. For example, an instructor said, “I expect students to use it to avoid doing the work for the class, and I am worried that I won’t be able to tell the difference between work they produce themselves and work generated by the AI” (P85). Instructors often wrote about countermeasures to these challenges. For example, an instructor said, “The primary challenge to me seems to be students submitting work completed by the AI. I would tailor assignments so it must be personalized and difficult for the AI to complete and hopefully learn ways to use AI in the classroom” (P299).
While business communication instructors believe teaching should change for the AI Age, the scale of the anticipated change has created anxiety among many business communication professors. About 47 percent of those in this study feel nervous or anxious about using it in class (see Table 2). Several commented they were relieved they would retire soon instead of facing this wave of change. For example, one instructor said, “Frankly I hope to retire rather than face this :)” (P74).
Anxiety Among Business Communication Instructors.
Note. Items were on a Likert scale from 1 = strongly disagree to 7 = strongly agree. “% Agree” refers to the percentage of respondents who selected one of the following items: 5 = somewhat agree; 6 = agree; or 7 = strongly agree.
Concerns About AI-Assisted Writing
Instructors hold many concerns about AI-assisted writing. In the quantitative portion of the study about concerns (see Table 3), majorities of instructors held concerns about all listed items. Instructors are concerned there will be more plagiarism and evaluation of work will be more challenging. They believe generative AI will lead to less critical thinking and creativity. They are concerned generative AI will lead to less perceived authenticity and credibility of communicators. Through the remainder of the section, we categorize these concerns in terms of accountability, agency, and authenticity and provide comments from instructors to illustrate these concerns (Table 3).
Concerns With AI-Assisted Writing.
Note. Items were on a Likert scale from 1 = strongly disagree to 7 = strongly agree. “% Agree” refers to the percentage of respondents who selected one of the following items: 5 = somewhat agree; 6 = agree; or 7 = strongly agree.
Concerns about accountability and agency
More than any other issue, instructors are concerned about accountability. They are fundamentally concerned with students taking responsibility for the appropriate development of quality content. This begins with students’ willingness to abide by standards of academic integrity. About 83% believe generative AI will lead to more plagiarism. Instructors also worry that generative AI can lead to less critical thinking and creative thinking, which are critical for quality content development. Over three quarters of instructors said generative AI will lead to less critical thinking (77%) and less creativity in writing (75%). Also, many instructors worry that students will not take responsibility for inaccurate and biased information that is generated by AI. For many instructors, closely tied to the notion of taking responsibility for the development of quality content is the notion of agency. Instructors worry students will become overly dependent on AI, thus constraining their skill development and choices in the short and long run.
In open-ended comments about the challenges of AI-assisted writing, academic integrity was among the most common concerns (see Table 1). For example, an instructor mentioned: “The primary challenges are related to ensuring academic honesty and integrity. I will address this topic with my students in class and ask them to think about and discuss why it is problematic for AI to think and write for them” (P158).
Instructors also frequently discussed the potential loss of critical thinking and creativity. Some instructors see the loss of critical thinking as an extension of an ongoing trend. An instructor said, “There are countless new technologies and services that deter us from creative thinking skills in the wake of what Time Magazine dubbed ‘The Creativity Crisis.’ This is yet another issue that leads us to atrophy of our abilities” (P58). Some instructors foresee the problem may become more significant as people use generative AI more. For example, an instructor said, “As we utilize AI more, our critical thinking skills will diminish. I also worry our jobs might become obsolete if the technology becomes more nuanced” (P133).
Closely related to a concern with the loss of critical thinking is a concern with less development of writing skills. Instructors often adopt a view similar to the following statement: “Writing is a process. Writing to learn = learning to write” (P62). An instructor elaborated this idea: “A primary challenge is that it removes the process of writing for students who use it. It is a process to plan or draft—and then write—and then revise. Writing is not a one-time transaction” (P110).
Many instructors worry that students will not take responsibility or not notice that AI-generated information is often inaccurate or biased. They worry about AI-generated content that contains “false, biased, or outdated information” (P176). Put simply, some instructors worry that AI “will manipulate the human thought process and generate deception of thought” (P176).
Some instructors connect the loss of critical thinking and creativity with a loss in agency. One instructor said, “It restricts student agency. . . . Students are already so dependent on secondary sources, and tend to plagiarize or ‘paraphrase’—we don’t need something like ChatGPT. Let brains think!!” (P146) Similarly, instructors worry that students will lose curiosity and discovery skills. One instructor said, “I am concerned that my students will rely too heavily on information and data ‘discovered’ by ChatGPT. I am concerned that ChatGPT will reduce students’ ability to conduct effective research and think critically” (P232). An instructor commented, “Artificial Intelligence has been a blessing and a curse. It makes our lives easier, but it also makes us less skilled and capable” (P57). Some instructors emphasized that AI could fundamentally alter human potential. One instructor explained, “I worry that students will adopt characterless, functional writing habits that do not allow them to think through problems on their own or with teams. AI ‘thinking’ can be limited, whereas the human imagination is unlimited” (P247).
Concerns about authenticity
Instructors expressed concern about AI-assisted writing leading to less authenticity. In the quantitative survey items, roughly 73 percent believe that AI-assisted writing will lead to lower perceived authenticity. About 62 percent believe it will lead to lower perceived credibility of communicators. Generally, instructors explain in open-ended comments that a range of issues affect authenticity, including developing one’s own voice, personalizing a message for others, showing sincerity, and otherwise inserting the human element into written communication.
Instructors want their students to develop their own voices and styles. An instructor commented, “It may impede students from doing the admittedly hard work of . . . developing their own authentic voice and ideas” (P23). Another instructor commented on the AI-assisted writing environment: “The primary function may emerge to generate writing products and documents versus a tool to sketch out and creatively develop one’s unique voice, style, tone, and quality of writing for a targeted audience” (P59). Instructors worry the loss of voice will result in “writing all sounds alike” (P44).
Instructors view relationship building and audience orientation as important elements of authenticity. One instructor stated this view in terms of soft skills: “Another challenge will be ensuring that my students cultivate the soft skills like empathy, audience analysis, and relationship-building without learning to write and assess an audience as a means for doing that” (P323). Some instructors foresee the depersonalization of communication as the biggest challenge. One instructor said, “AI, such as ChatGPT, will . . . further deteriorate writing/rhetorical skills of many, if not most, business professionals. The depersonalization of communication will further alienate employees and customers, leading to more dissatisfaction in the workplace” (P132). Others adopt a view of the opportunity: “It also might provide us with an opportunity to talk about what computers and technology CANNOT do like understand context, empathize, and analyze and anticipate audience emotion” (P323).
Teaching AI Literacy
Although generative AI (i.e., ChatGPT) has only been widely available for less than six months, many business communication instructors are already teaching about AI-assisted writing. Others are actively planning to teach about AI-assisted writing. Instructors tend to focus on integrating AI-assisted writing in ways that focus on broad principles of effective writing. They also focus on using AI-assisted writing in the beginning stages of writing for idea generation and early drafting with a vision of extensive revisions. Fewer instructors mention using AI-assisted writing for editing and evaluation. Many of the strategies that instructors mention can address key capabilities of AI-assisting writing. As a result, we allude to application, accountability, authenticity, and agency throughout this section.
For teaching about AI-assisted writing, the most common teaching approaches involve evaluating ChatGPT output and/or making side-by-side comparisons of ChatGPT output with that of students or other people. This process generally involves elevating the conversation to broad principles of effective writing. One instructor explained how evaluating the output provides an opportunity to explore genre analysis (which we consider accountability, taking responsibility for high-quality content) and an audience orientation (which we consider connected to authentic communication): It’s [ChatGPT] very good at coming up with decent generic writing, seems to have a pseudo-“critical” tone, is excellent at recycling information and putting it together in a reasonable way. However, it lacks depth and a bit of variety. . . . Consequently, I think it could be a great starting point for discussion, using it to help students do genre analysis on boilerplate-type texts or to identify features of what ChatGPT thinks makes a text appropriate for a specific audience. (P183)
Another instructor noted they will use this comparison approach to encourage student agency: There is no stopping these tools from being widely used in the workplace, so introducing these tools in the classroom helps prepare students for the real world. I will have students complete assignments where they can see, side-by-side, their “scratch” work and ChatGPT work and decide for themselves how they would use it in the future. (P216)
While instructors most often emphasized the nontechnical matters of accountability, authenticity, and agency, some are also focusing on technical expertise (i.e., application) to interact effectively with generative AI. For example, an instructor said the following: Students need instruction on how to properly use the tool. You have to train them to use their critical thinking skills to properly “program” ChatGPT to produce the output desired. If you go into ChatGPT without being prepared—audience, purpose, goals, etc.,—your message will fail. . . . As instructors, our job is to teach them the pros and cons of using it, relying too heavily on it, not being trained in how to effectively design prompts for it. (P143)
This instructor emphasized issues of accountability (critical thinking), authenticity (audience orientation), and agency (avoiding overreliance). This instructor also emphasized the application aspect of the tool by alluding to the skill of developing effective prompts or questions while interacting with ChatGPT. This theme is common among these instructors: issues of accountability, authenticity, and agency drive application considerations.
Many instructors explain that generative AI can be particularly useful in the idea development and/or drafting stage of writing. For example, an instructor said: I think ChatGPT could be useful for helping students overcome the blank-page block and to get good starting ideas for a project. It could also give them a sense of what a good organizational structure could look like early in their drafting. . . . I can also imagine using ChatGPT myself as a starting place for generating model texts, since it’s tough to find real-world samples of business documents to use for instructional purposes. (P324)
Similarly, another instructor mentioned that in this early stage of writing, generative AI could help groups as well: “ChatGPT provides excellent suggestions and could be a great aid in providing guidance and examples for students. I could see myself suggesting that students use it as a launchpad for their team research proposals because many struggle when they first begin this assignment” (P192).
Instructors also emphasized ethics. For example, an instructor explained: “I see ethics as the primary challenge. I’ve begun discussions with my students in classes about the realities, advantages/disadvantages, and the ethics of using AI assisted software when they are supposed to be demonstrating their own competency in . . . writing” (P123). We characterize ethics as an aspect of accountability for content development.
One common theme about the teaching strategies adopted by instructors is that they involve higher-order thinking skills that are often facilitated through discussion and experiential learning. For example, an instructor commented, “Often, beginning a writing assignment is the challenge. ChatGPT eliminates that challenge and makes it easier to move to analyzing and evaluating. It may help with moving toward higher levels of thinking in Bloom’s Taxonomy” (P206). Another instructor emphasized this point: I have already assigned ChatGPT use for my class. We will ask the AI to draft cover letters and then critique them for improvements. This is simply asking students to “think a level up” on Bloom’s Taxonomy. . . . I will also let students use AI to draft a speech because what I want to teach is how to deliver a good speech. We’ve all sat through those speeches by well-meaning executives who THINK they can give a good speech but can’t. I am teaching how to present a message in a way that the audience will listen. AI doesn’t conflict with that. (P307)
Generally, many instructors who adopt this view of higher-order thinking adopt a view that can be characterized in this way: “[The opportunity is] writing alongside Chat GPT and training students to be the brain in the equation” (P303).
Although instructors focused primarily on how to teach about AI-assisted writing in class, some are considering ways to use generative AI to develop lesson plans and aid in assessment. One instructor commented, “To use it while preparing for the classes, as a tool that can help to create lesson plans, presentation outlines, give ideas concerning student projects, etc.” (P169). Another instructor commented, “I know that it can help instructors develop assignments, rubrics, and even syllabi. It’s also a useful tool for brainstorming and creating boilerplate text” (P264). Other instructors were eager to experiment with ChatGPT to evaluate student work. For example, one instructor said, “To review writing, especially for continuity, readability, and grammar/punctuation errors” (P2). Another said, “Today I saw someone had used ChatGPT to apply a rubric to evaluate a writing. I was giddy with the thought of the labor saving possibilities” (P54). Interestingly, some instructors noted the potential for instructors to use generative AI to evaluate work that students accomplished with the aid of generative AI. For example, an instructor commented, “I’m wary of having students use it to do their writing and then me use it to do the grading–just bots talking to bots.”
Policy-Making and Equity Considerations for AI-Assisted Writing
Many instructors in our study are eager for policy and guidelines as they navigate the AI Age. We created a forced-choice survey item for instructors to identify their preferred policy-making preference (see Table 4). The top two choices were for policy developed at the university/school level (33.7%) or the departmental level (24.3%). Other instructors preferred policy driven by individual instructors (19.8%), and some instructors want generative AI banned (12.7%). This forced-choice question reveals a wide range of policy preferences.
Policy-Making Preferences for AI-Assisted Writing.
In open comments, instructors similarly stated many policy-making preferences. Most of those who commented, however, did seek at least some centralized guidance and set of parameters. For example, one instructor commented: “I am planning to bring this up in my next faculty meeting to brainstorm the department/program-level policy at least. But I think it would be good to have a university-level policy. I think we all are in a wait-and-see situation right now though” (P89). Another instructor stated: Given that automated writing will come to stay, students/learners need to know how to use AI-based applications or tools, and educational institutions need to factor this into courses of learning. We need to persuade students to learn the skills on their own before going for automated tools. Schools and colleges need to regulate and limit the use of automated writing tools as a matter of policy, as this is required across courses and instructors. (P337)
Some professors, however, worried about rigid, rushed, and short-sighted policy making. For example, an instructor commented, “[The challenge is] helping others to see the benefits. In the Liberal Arts college where I work, many of my colleagues feel the need to ‘police’ students and resist change” (P104). Another instructor commented, “The biggest challenges will stem from trying to fight against it via heavy-handed policy due to plagiarism fears” (P204). “The primary challenge is the fear, mostly unjustified, faculty and administrators have of ChatGPT” (P213).
Those who want to ban the use of generative AI tend to hold strong feelings, largely out of their desire for academic integrity and student growth. They see AI-assisted writing as incompatible with the development of writing skills and critical thinking. For example, an instructor offered this view: ChatGPT—and let’s be honest; there will be others—is the final nail in the coffin of communications learning. . . . Being spoon-fed what they need to pass a writing course is detrimental to any chances that they will learn the skills they need to use when access to such tools is limited and they are on their own. I plan to aggressively prohibit the use of this tool in my classroom, to the point that I will require students to do their writing assignments while in class. (P270)
Interestingly, instructors hold competing views about whether AI-assisted writing would be more or less equitable. Many instructors commented that it would be more equitable for second-language learners, poor writers, and those with learning challenges. For example, an instructor said, “There is a great opportunity to help students for whom English is a 2nd/3rd language to improve their writing or command of the language” (P144). Another instructor said, “It can act as a good resource for reading, understanding and implementing writing strategies for students who are weak in communications” (P147). Another instructor believed the tools could create more equity for individuals with neurodiverse backgrounds: I have a new perspective on this topic than if you had asked me a year ago. My 8-year-old daughter was diagnosed with dyslexia a year ago. She has brilliant ideas, but she struggles to express them in writing. I can see this technology alleviating some of the over-emphasis that we (our society) place on the WRITING part as opposed to the THINKING part. (P24)
Other instructors, however, worried that the tools would disproportionately harm individuals of certain groups, including those who are poor writers. For example, and instructor commented: “Students will use this as an easy way to complete shorter assignments. The students who use this to complete assignment may be the students with lower writing skills already, who most need the skills taught in our classes” (P67). Another instructor explained, “It will end up being just one more technology divide. The weaker students will tend to gravitate to it; the stronger students know better” (P252).
Perspectives About Future Use of AI-Assisted Writing in the Workplace
Business communication instructors often attempt to align their course content with the knowledge, skills, and attitudes necessary for successful employment. As a result, we wanted to understand the degree to which they anticipate adoption among business professionals. Similarly, we wanted to understand their general views about what they view as the benefits of AI-assisted writing.
Based on items from Venkatesh’ et al. (2003, 2012) technology model, most instructors believe AI-assisted writing will be useful, easy to use, encouraged for use by managers, and widely adopted in the workplace (see Table 5). The most important factor in adoption is typically considered usefulness. Roughly 80 percent of instructors believe it will be useful in the workplace, and nearly 85 percent believe it will enable professionals to complete tasks more quickly. Roughly 80 percent believe it will be easy to use. Just over 60 percent of instructors believe AI-assisted writing will be used by most professionals in the near future.
Expected Use of AI-Assisted Writing in the Workplace in the Near Future.
Note. Now think about your expectations of how AI-assisted writing will be used in the workplace in the future. In the near future (within years), AI-assisted writing will . . .
In open comments, instructors reinforced these views that AI-assisted writing will be widely used in the workplace. For example, one instructor commented: Professionals writing copy will need to know how to write with it. They’ll be expected to use it to speed up brainstorming, outline-making, and drafting. I may use it to experiment with social media post copy and blog headlines for my professional writing students. We will likely contrast ChatGPT produced work with what they write for blogs. I’m eager to see what we discover. (P243)
Like this instructor, many followed up their statements about how professionals are or will generative AI with comments about how to align training in the classroom with future needs. In this case, the instructor emphasized using generative AI to explore use with professional blogs. Other instructors emphasized the importance of business communication instructors framing the discussion. For example, an instructor said, “If this is the communication wave of the future in the workplace, then bus comm instructors need to get knowledgeable and become the drivers of it” (P103).
Regarding benefits, roughly 63% believe AI-assisted writing will be more efficient, and about 56% believe it will create opportunities to generate ideas for writing. Although most instructors believe AI-assisted writing will be common in the workplace and generally accept that it will lead to more efficient writing and opportunities for idea generation, less than half of instructors agree on other types of potential benefits listed in our survey. For example, far less than half agree that AI-assisted writing will lead to writing better aligned with audience needs (38.7%), contain more critical thinking (21.7%), or help them grade or asses student work (18.1%) (Table 6).
Perceived Benefits of AI-Assisted Writing.
Note. Items were on a Likert scale from 1 = strongly disagree to 7 = strongly agree. “% Agree” refers to the percentage of respondents who selected one of the following items: 5 = somewhat agree; 6 = agree; or 7 = strongly agree.
Discussion and Recommendations
We entered this project with the assumption that the business communication community could uniquely provide perspectives about how students can succeed and thrive in the AI Age. Business communication instructors develop their expertise with applied skills in mind and align their teaching with a growth mentality, seeking to support students to communicate more strategically, influentially, and ethically in a wide range of professional situations across their careers. The consensus view of instructors in this study is that AI-assisted writing specifically and generative AI broadly will be widely used for business communication. We agree with experts who believe the AI Age is emerging (Kissinger et al., 2023). The widespread use of generative AI necessitates significant changes to learning and teaching business communication. Collectively, instructors suggest that professionals and students will need AI literacy—involving application, authenticity, accountability, and agency—to communicate effectively. To enable instructors to provide students with adequate preparation for the AI Age, administrators will need to provide resources and a supportive environment for instructors.
Developing AI Literacy as Communicators
Assuming generative AI will lead to profound changes in the workplace and the teaching environment (as perceived by the vast majority of instructors in this study), we believe all students should become managers of AI. We suggest the collective wisdom produced by the instructors in this study can be characterized as AI literacy and involves four sets of capabilities: application, authenticity, accountability, and agency. A fundamental outcome of university programs should be for students to develop AI literacy (see Figure 1; a set of guiding questions for each aspect of AI literacy is presented in Table 7). Business communication courses are among the most important courses to develop it given the applied focus on communicating in the workplace.
Capabilities of AI literacy for communication.
Guiding Questions for Effective Management of AI in Business Communication.
Fundamentally, students should understand the application of AI to their school and work activities. They will need an understanding of various generative AI tools, including their strengths and weaknesses. They should understand how to align the capabilities of these tools with the tasks at hand. The widespread adoption of ChatGPT implies these tools are easy to use, at least in a rudimentary manner. Many of the online demos and use cases of these tools have focused on how to use generative AI more effectively. For example, Mollick (2023) has produced many demos of how to adjust the instructions, queries, or prompts to achieve better results. One writing coach has suggested approaches to adjust tone with ChatGPT (i.e., use commands like “more optimistic,” “more serious,” or “more lighthearted) (Bush, 2023b) and more goal-oriented (i.e., use commands like “more persuasive,” “more descriptive,” “more action-oriented,” or “more concise” (Bush, 2023a). Much of the online commentary involves improving the application of generative AI.
On a more profound level, current and future professionals will need to develop AI literacy in authenticity, accountability, and agency. Business communication instructors are uniquely positioned to support the growth of current and future professionals in these areas because they are compatible with traditional approaches to business communication applied to emerging forms of communication. Authenticity involves focus on genuine communication and prioritizing the human element. Even with a glimpse of the emerging capabilities of generative AI, business communication instructors in this study clearly emphasize that communicators need to personalize and tailor messages with their own voice and the unique needs and wants of their audiences. Aside from this study, other emerging research shows people believe that AI-mediated communication is less authentic, even when communicators disclose they used AI (Glikson & Asscher, 2023). Thus, the ability to communicate authentically will require much more attention and care in the AI Age.
Accountability involves taking responsibility for the accuracy and appropriateness of AI-generated content and using generative AI in a fair and equitable manner. Professionals will need to develop higher information literacy, better discernment to separate information from misinformation, and an undeviating commitment to reliability in communication. Historically, people tend to judge AI mistakes more harshly than mistakes made by people (Dietvorst et al., 2015), yet as professionals use AI to communicate in their own names, it is likely that others will not give any leeway in terms of inaccuracies, misinformation, and other forms of unreliable communication.
Agency involves people retaining control to make their own choices. A focus on human agency is particularly important in the context of AI-influenced communication (Floridi, 2023; Kang & Lou, 2022). A rich tradition of human-in-the-loop research explores the ways in which people and AI can interact to achieve better outcomes while ensuring there is always human input and control (Zanzotto, 2019). Getchell et al. (2022) theorized about various roles AI can take in business communication, including a tool, an assistant, a monitor, a coach, and a teammate. Their work suggested the teammate role was unlikely to have any realistic application in the near future, yet the release of ChatGPT has resulted in scholars already debating whether ChatGPT should be considered a coauthor on published work (Biswas, 2023). The widespread availability of generative AI will likely allow AI roles that could lead to less control for professionals. Increasingly, business communication instructors will need to address how to maintain human agency in these situations. We consider the comments of instructors in this study largely aligned with the following statement from former Google CEO Eric Schmidt, former U.S. Secretary of State Henry Kissinger, and MIT professor Daniel Huttenlocher: Education in particular will need to adapt. A dialectical pedagogy that uses generative AI may enable speedier and more-individualized learning than has been possible in the past. Teachers should teach new skills, including responsible modes of human-machine interlocution. Fundamentally, our educational and professional systems must preserve a vision of humans as moral, psychological and strategic creatures uniquely capable of rendering holistic judgments. (Kissinger et al., 2023)
Supporting Students and Instructors in Their AI Journeys
We believe that a fundamental mission of higher education should be to prepare students for work in the AI Age. While instructors in our study collectively identified important elements of AI literacy, they also conveyed in many ways the challenges and constraints of adequately teaching AI literacy addressing these challenges on their own. Based on the challenges and constraints they raised, we offer our views on how to help students develop AI literacy through the purposeful and aligned efforts of instructors and administrators. By keeping students at the center of curriculum planning for AI literacy, instructors and administrators can prioritize their shared efforts (see Figure 2).
Placing students at the center of training for the AI Age.
Our study emerged from the viewpoints of instructors, who directly engage students in their development of AI literacy. We believe this transition will require a lot of effort and resources. As a result, we encourage a community-driven, administrator-supported approach to putting these recommendations into action. Instructors should consider the following recommendations:
Develop AI literacy
Instructors will need to continuously experiment with generative AI and evaluate its usefulness for various professional goals. They need the freedom, resources, and training to gain wide exposure to how contemporary professionals are using generative AI for workplace communication. They will need to constantly increase their expertise in each area of AI Literacy. While this recommendation may sound daunting to some, we encourage all communication instructors to recognize that developing AI-specific expertise in authenticity, accountability, and agency are extensions of their existing expertise in business communication.
Oversee creative and innovative learning and teaching
Generative AI will significantly impact the goals, priorities, processes, and outcomes of business education (including business communication courses). Even in the early stages of widely available generative AI, instructors in this study largely envision more focus on higher-order learning with more robust class discussions, experimentation with emerging technologies, more experiential and project-based learning, and new forms of evaluation. Most, but certainly not all, envision more time allocated to editing and revision, with less time allocated to first drafts and routine content collection. We encourage instructors to embrace new ways of teaching business communication without sacrificing the core elements of business communication, which remain timeless.
Operate in communities of practice
We believe the value of community is more essential in the AI Age. In some ways, the pandemic may have supported a movement in this direction in that many schools allowed more flexibility and emotional support for faculty, staff, and students. In the AI Age, instructors need supportive communities of practice to manage uncertainty and anxiety and develop their abilities to teach effectively. We also believe community is important so that new norms and practices are shared and negotiated with competing views valued and even celebrated. We encourage all instructors to tap into communities so they can take this AI journey communally (to manage well-being) and collaboratively (to manage performance).
These communities may include professional associations (such as the Association for Business Communication), university AI interest groups, department-level working groups, or many other groups. We urge instructors to avoid isolated AI journeys.
Administrators play a critical role in supporting instructors on their AI journeys and monitoring the regulatory environment. Instructors will need to rapidly develop these skills themselves. This is particularly challenging given the intense time constraints they operate in and the limited resources they have. Our view is that instructors need to actively experiment with the tools and frequently engage with current professionals who are daily developing new uses of the tools. The traditional timeline of higher education that relies on slowly developed, peer-reviewed research is likely too slow to benefit instructors in their AI journeys. Also, instructors indicate the uncertainty and anxiety in this environment. For them to experiment and innovate, they will need to engage in a variety of trial-and-error approaches to teaching. Specifically, administrators should consider the following recommendations that are supported in a vision of innovation and an environment of psychological safety:
Solicit cross-disciplinary input
Administrators should seek input from instructors with many types of expertise. In business schools, there are already indications that generative AI will significantly impact coursework and writing in finance (Dowling & Lucey, 2023), accounting (Alshurafat, 2023), marketing (Wertz, 2023), operations (Terwiesch, 2023), and specific industries (e.g., Patel & Lam, 2023). Administrators should particularly seek out the expertise of tech-savvy communication instructors who can help them understand the implications of generative AI for leadership, management, and communication.
Set community-based vision and policy
A recent study of early adopters of AI-based communication technologies showed that given the uncertain and constantly changing nature of AI, a social contracts approach to policy making is ideal (Cardon et al., 2021). Administrators should develop deep awareness of how policies support innovation while also protecting employees. By relying on their communities to understand these issues, they can develop effective policies that will benefit students, staff, and faculty. Several of us are currently involved in initiatives at our own institutions and have talked to practitioners who work in units that are experimenting with these tools. Our initial observations are that building community-based vision and policy requires a laboratory-type approach in which instructors and/or employees across functions and units experiment with the same uses or applications of generative AI. For example, each member of the community might generate an analysis of a particular business case through the use of ChatGPT. This approach creates common experiences from which to evaluate the benefits, weaknesses, and implications of generative AI in a structured manner.
Provide resources for innovation
Participants in this study repeatedly, explicitly and implicitly, revealed the challenges associated with helping students develop AI literacy. Administrators should recognize that instructors will need substantially more resources, including training, access to communities of practice, and, most importantly, time.
Pause institutional constraints and build culture
Instructors in this study generally view generative AI as a significant disruptor. To innovate, instructors need resources, encouragement, and psychological safety. Administrators will need to consider how higher education culture and structures may hinder instructor-driven innovation. For example, excessive reliance on student evaluations as a measure of teaching effectiveness may counter the trial-and-error approach necessary to innovate. Instructors clearly suggested that teaching and evaluation will require more higher-order learning (i.e., Blooms taxonomy) that is facilitated by more robust class discussions, more experiential and project-based learning, new and more often updated assignment prompts, and more constant professional development to stay current with rapidly evolving technologies. We consider large class sizes and high teaching loads as incompatible with these changes. Administrators will need to carefully consider how to reallocate other resources in higher education to allow smaller teaching loads and smaller class sizes in which instructors can deliver more personalized learning. We believe that a key component of a school’s value proposition in the AI age will be to deliver more personalized interactions between instructors and students.
Limitations and Future Research
We acknowledge several limitations to this study. This study captures attitudes and reactions at an early stage of generative AI, including AI-assisted writing. AI-assisted writing is still new to most users, and new and improved AI tools will continue to be offered from major software vendors, including but not limited to, Microsoft, Google, and Facebook (Pichai, 2023; Schechner, 2023). Over the upcoming years, it is difficult to predict the many ways in which these tools will be used in business communication. As the use of generative AI is connected to new forms of data, such as employee information and prior communication, new issues of privacy and authenticity will likely increase. Surely, regulation (e.g., GDPR; corporate policies, especially in regulated industries) will influence the use of generative AI. This study is also limited by capturing only the views of instructors. Within the higher education system, research involving administrators and students is also needed. In addition, this study does not capture the experiences of practitioners, which is a critical benchmark from which to develop curricular content for higher education.
Opportunities for additional research are abundant. Each aspect of AI literacy—application, authenticity, accountability, and agency—has many directions for additional inquiry. The guiding questions listed in Table 7 not only serve as important discussion questions and standards for students, they also support many direction in research. The landscape of research about AI is broad, so it is likely that this framework of AI literacy will need to be elaborated, expanded, and interrogated in practice and in research. Yet, little research is available about business communication practice and pedagogy (Getchell et al., 2022). We urge business communication scholars to take a leading role in exploring how generative AI influences the practice and instruction of business communication.
Footnotes
Appendix A: Survey
Appendix B: Coding of Open-Ended Responses
Code Descriptions and Examples.
| Code | Description / Examples |
|---|---|
| Teaching needs to change | Explicit statements about the need for teaching to change due to ChatGPT and other generative AI tools. Examples include the following: • “The whole concept of learning should be changed. The future is here.” (P129) • “The speed with which ChatGPT is evolving. Instruction for this tool is already way behind its actual usage. Teachers need to learn more about ChatGPT, use it themselves, and figure out creative ways to teach it.” (P166) |
| Resistance to change in teaching | Explicit statements about anxiety about changing teaching and learning, not having resources or training to change, not having enough time to prepare for change, or not wanting to address the change (e.g., to ban ChatGPT). Examples include the following: • “I feel that technology is advancing at warp speed, and I worry I don’t have the time to create new assignments or discussions that can show students how to use it in their professional lives as a tool but not as a replacement for communication skills development.” (P141) • “ChatGPT might provide professionals with the ability to properly organize an essay/message. I do not plan to take advantage of ChatGPT as it’s a program that is going to cause our learners to continue to remain ignorant and not think for themselves.” (P260) |
| Evaluation of student work will be more difficult | Distinguishing the work of students versus the work of AI will be more challenging. Examples include the following: • “It would be difficult to catch cheating.” (P140) • “Professors in fields across campus have students write because it is one of the best ways to internalize knowledge. With AI assisted writing that is practically untraceable, we are losing one of the key indicators of one’s ability to think critically and communicate effectively.” (P215) |
| Academic Dishonesty from Students | Plagiarism, cheating, and any unauthorized use of AI-assisted writing. Examples include the following: • “Business students are already notorious for cheating, students since the pandemic appear to be more likely to cheat, and the dearth of students over the next decade will make it more likely that cheating will be overlooked. So, yeah. Cheating.” (P125) • “I think plagiarism will rise exponentially. I also believe that students will become even less engaged in the act of learning how to write properly.” (P224) |
| Less critical thinking in writing | Reduced critical thinking in writing due to generative AI. Examples include the following: • “Wholesale use of AI tends to shortcut the critical thinking process that develops confident writers and thinkers.” (P29) • “It takes away human creativity, learning, deductive reasoning, critical thinking, and the ability to process thought. While the ChatGPT platform has a wide range of applications related to text generation, outputs should be evaluated critically and used with caution.” (P176) |
| Unreliable writing | Unreliable writing due to generative AI. Unreliability may be due to inaccurate information, fake or fabricated news, and bias. Examples include the following: • “Garbage in / Garbage out. The content generated is only as good as what is out there on the World wide Web.” (P82) • “Another challenge is how to check for accuracy regarding content. AI is not perfect in content. ChatGPT said that Hillary Clinton was the first woman president from 2017 to 2025.” (P260) |
| Less development of writing skills | Concern that students will not develop fundamental skills needed for effective writing. Examples include the following: • “Students won’t actually learn how to write; they will rely on the AI and will not even review the document to determine if it meets the criteria of the assignment or the needs of the audience. I will address this by discussing how students need to do the work in order to learn how to do the necessary tasks, showing how this skill will make them better employees and stronger candidates for jobs and promotions.” (P53) • “Writing is a process. Writing to learn = learning to write. ChatGPT will seriously change the game we play with our students about teaching writing as a process. Banning AI-generated text seems the only option.” (P62) |
| Less authentic writing | Reduced authenticity in writing due to generative AI. Reduced authenticity may be due to less individual voice and originality, less personalization of messages, a lack of audience orientation, and other loss of the human element. Examples include the following: • “AI, such as ChatGPT, will further reduce resources and focus for writing instruction. It will further deteriorate writing/rhetorical skills of many, if not most, business professionals. The depersonalization of communication will further alienate employees and customers, leading to more dissatisfaction in the workplace.” (P132) • “Students may lean on ChatGPT when more creative thinking and more audience-tailoring would improve their results.” (P139) |
| Lack of guiding policy surrounding AI | Explicit statements about desiring more clarity in AI policy. Examples include the following: • “Acknowledging the benefits and pitfalls of Chat GPT in the classroom may help curb dependency and incorporating a policy against using it to write their papers (much like plagiarism) could deter its use.” (P20) • “The primary challenge is to integrate it into my course curricula. I am planning to bring this up in my next faculty meeting to brainstorm the department/program-level policy at least. But I think it would be good to have a university-level policy. I think we all are in a wait-and-see situation right now though.” (P89) |
| Jobs will be eliminated | Explicit statements about job losses for writing professionals and for professors due to generative AI. Examples include the following: • “It will be increasingly difficult to convince others of the value of learning to write. I am also concerned about how ChatGPT and similar platforms will enable users to lack of [sic] critical thinking skills. Finally, I’m not sure how to prepare students for the workforce—what jobs will be left? I don’t have answers for these issues.” (P264) • “Marketing and communication jobs have already been impacted by increased precarity, overseas suppliers and corporate downsizing. . . . This will further decimate that sector.” (P280) |
| Equity will decrease | • “For novice writers, ChatGPT may become a copout for them, thus, missing opportunities to grow and improve as writers.” (P96) • “Good students will improve by using it, because they can critically reflect the outcome. Bad students will not benefit and will probably fall even more behind.” (P137) |
| Discuss and evaluate generative AI in the context of effective writing principles | Actively engaging students with discussions and demos of AI-assisted writing with a focus on broad principles of effective writing. Examples include the following: • “One could use AI text to compare with authentic text to point out noticeable differences and highlight both benefits and shortcomings of AI writing. (P37) • “Students need instruction on how to properly use the tool. You have to train them to use their critical thinking skills to properly ‘program’ ChatGPT to produce the output desired. If you go into ChatGPT without being prepared—audience, purpose, goal, etc.,—your message will fail. The biggest problem with AI is when people want to ‘hide’ it or block it. Students already know about the ‘elephant in the room.’ As instructors, our job is to teach them the pros and cons of using it, relying too heavily on it, not being trained in how to effectively design prompts for it.” (P143) |
| Use generative AI for a first draft and to generate ideas for specific documents | Asking and even encouraging students to experiment with generative AI, particularly in the prewriting stage. This focuses on what students do with generative AI as part of their writing for specific documents. Examples include the following: • “ChatGPT can be used for a first draft. I, at this point, don’t see a problem with a student using it to help generate a first draft. I have no problem if my institution allows it or even encourages it to help generate a first draft.” (P8) • “ChatGPT could be used to provide examples of the various ways someone could respond to a given query, assignment, etc. I can see how an instructor could ask students to use ChatGPT for an initial draft and then ask students to edit and revise the draft ChatGPT provided.” (P131) |
| Integrate generative AI into assignments | Creating assignments that direct students to use generative AI, or using generative AI to create prompts and assignments. Examples include the following: • “Assigning a project that involves using ChatGPT could go far in helping students understand how to incorporate it as a resource to minimize reliance.” (P20) • “I know that it can help instructors develop assignments, rubrics, and even syllabi. It’s also a useful tool for brainstorming and creating boilerplate text.” (P264) |
| Use generative AI to grade and evaluate | Using generative AI to assist in grading and evaluating student work. Examples include the following: • “Today I saw someone had used ChatGPT to apply a rubric to evaluate a writing. I was giddy with the thought of the labor saving possibilities.” (P54) • “I have heard ‘rumors’ that it could facilitate the instructor-giving-feedback process.” (P82) |
| Writing will become more efficient. | Instances in which instructors believe generative AI will make writing in the workplace more efficient. Examples include the following: • “Efficiency, structure, and prompts to overcome writing blocks. As with any AI, the goal is to optimize, harness, benchmark, and provide objectivity that complements and challenges the human touch in feedback and production.” (P59) • “I could see it being helpful in the workplace for routine tasks. For example, automated email type situations to respond to common customer questions. I already use my own email templates to share with my GTAs to answer commonly asked questions in the large class I teach.” (P67) |
| Writing will be supported with ideas and knowledge. | Instances in which instructors believe generative AI will support idea development for professional writing. Examples include the following: • “For users, it is great benefit that they collect other’s ideas and experiences through internet and extend their new thoughts and skills.” (P79) • “I see some of the opportunities offered by ChatGPT as a sort of how Wikipedia works. Being able to search and find information and restate a sentence or piece in a different style or other transformations.” (P124) |
| Accessibility and equity will increase | Instances in which instructors believe generative AI will reduce disadvantages for some writers. Examples include the following: • “It could be used as a teaching tool. It might help ESL students.” (P4) • “ChatGPT could bring more equity to writing. Students who are less prepared, by no fault of their own, can use this tool to help them further develop their writing skills and make their ideas shine. ChatGPT could also drive more innovative thinking.” (P326) |
| No benefits or opportunities | Instances when instructors state clearly and definitively there are no benefits to AI-assisted writing. Examples include the following: • “I don’t see any opportunities presented. I think ChatGPT is the lazy way to complete work. Perhaps I don’t know enough about it to pinpoint any benefits. I think it’s a completely unethical tool.” (P212) • “I see no primary opportunities for students learning this program and I hope my university bans it.” (P224) |
| Not sure of benefits | Instances when instructors state they are unsure what the benefits of AI-assisted writing might be. Examples include the following: • “I’m not sure, but i think we need to find ways to embrace this tool. Fighting it will be a losing battle.” (P162) • “Not sure - I would think that students could potentially use it to write cover letters where they put in the position, information about themselves, and key skills. However, I think it still would leave out the specifics that only the student would know.” (P330) |
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
