Sage Journals: Discover world-class research

Abstract

The emergence of generative AI (GenAI) has illustrated that higher education needs to adapt to the technology. Its speed of evolution requires that we adequately prepare students for an ever-changing landscape. Toward achieving that aim, we draw on the concept of interpretive flexibility, where the interpretations, uses, and outcomes of a new technology can differ and evolve over time, often with dominant stakeholders controlling the process. To engage marketing students in this process, we propose that they be presented with these diverse interpretations now as part of GenAI literacy. Specifically, we offer three small-scale pedagogical interventions designed to address this urgent need. Given the newness of GenAI, our interventions are designed to be infused into existing marketing instruction, instead of requiring a redesign of a curriculum. With each intervention, students not only significantly decrease their confidence in the accuracy of what GenAI produces but also see reasons to examine the implications of it. Both these outcomes, we suggest, could help to maintain interpretive flexibility required to properly respond to and guide the technology as its uses, impacts, and evolution become evident. We encourage educators to prioritize a comprehensive notion of GenAI literacy in their pedagogy to maintain interpretive flexibility.

Keywords

Generative AI pedagogy marketing education AI literacy interpretive flexibility technology in the classroom

Generative AI (GenAI) exploded into the public conscience in November 2022, when OpenAI’s ChatGPT was released for mass consumption (Guidi, 2023). ChatGPT is an internet-based application of large language models to process natural language, along with a variety of other similar offerings such as Google’s Gemini and Anthropic’s Claude (Cook, 2024). GenAI is “a technology that (i) leverages deep learning models to (ii) generate human-like content (e.g., images, words) in response to (iii) complex and varied prompts (e.g., languages, instructions, questions)” (Lim et al., 2023, p. 20). GenAI has low-threshold requirements for use. For example, ChatGPT can be accessed for free using everyday language and requires only an internet connection (OpenAI, 2024). According to recent estimates, by 2032, GenAI is poised to be a $1.3 trillion market (Bloomberg Intelligence, 2023) and is expected to have widespread implications for marketing (e.g., marketing content, market research, and customer support; Korst & Puntoni, 2023; Kshetri et al., 2023). Therefore, we must urgently consider how to provide relevant training in marketing classrooms (Adiguzel et al., 2023; Ye et al., 2024) to prepare students for these future jobs (Domínguez Figaredo & Stoyanovich, 2023; Peres et al., 2023).

However, instead of merely teaching students how to use this emerging tool (i.e., providing technology training), marketing instructors can—and should—directly encourage interpretive flexibility (e.g., Meyer & Schulz-Schaeffer, 2006) around GenAI. Interpretive flexibility captures how technology is constructed and interpreted by people with the consequence that meanings, uses, and problems vary across stakeholders, that is, across “relevant social groups” (Meyer & Schulz-Schaeffer, 2006, p. 1). This variety of interpretations tends to stabilize over time as the technology develops (e.g., is modified to address problems) until a consensus—or “closure”—is reached (Pinch & Bijker, 1987). Importantly, closure may be determined by consensus across stakeholders, or by one or more dominant stakeholders (Bijker, 1987). As this social process of interpretive flexibility unfolds, different uses of a technology and the contexts in which they are situated result in a variety of consequences (Doherty et al., 2006), many unintended, unforeseen, and only revealing themselves in the longer term. Viard, Gornet, and Delarue (2024) have called for interpretive flexibility as a normative tool “to root data science and machine learning in its sociotechnical [vs. purely technical] nature” (p. 2), thus addressing implications for different stakeholders and society at large, and also to avoid further entrenching existing power structures.

Consequently, as GenAI evolves and is purposefully modified, students need to remain flexible in how they use, construct meaning, and understand the (unintended, multi-stakeholder) implications of the technology as these consequences materialize. This requires universities—and particularly business schools—to foster skills beyond technical abilities to shape socially-responsible leaders who remain open to these evolving uses and implications, with multiple stakeholders involved, including students, instructors, and university administrators. For educators, this means sustaining students’ interpretive flexibility of GenAI by fostering understanding of this technology, including what is possible and what could be different.

Instructors can do this through nurturing GenAI literacy. This aligns with the purpose of business education more broadly, as expressed, for example, in the Association to Advance Collegiate Schools of Business’ (AACSB, 2024) assertion that “[b]usiness schools play a pivotal role in cultivating desirable talent by equipping students with a comprehensive skill set that includes AI literacy” (n.p.). As a way forward, we propose focused interventions that can be utilized in marketing courses to foster GenAI literacy and facilitate sustained interpretive flexibility across diverse social groups. We suggest the use of small-scale interventions, which can increase student engagement especially when the intervention is novel (Dyson, 2008). Small-scale interventions are also in line with the suggestion by Rodriguez-Tejedo and Etayo (2024) who advocated adding “low-cost and easy-to-implement activities” (n.p.), circumventing the challenge of wider curriculum change in the short term, especially when many educators lack the power to influence wider university policies and curricula in a timely way. To demonstrate the usefulness and appropriateness of focused interventions to facilitate interpretive flexibility around GenAI, we created interventions for three different marketing courses at the undergraduate and graduate levels and engaged in pre- and post-activity assessment of their effects.

The contributions of this study are manifold. One, we bring together two concepts, that of GenAI literacy and interpretive flexibility, in the context of marketing classrooms. Two, we argue that GenAI literacy requires socio-ethical considerations be embedded in the building blocks of GenAI literacy (understanding, usage, evaluation). Three, we show that even small-scale interventions can lead to improved GenAI literacy. Consequently, we position marketing education—and the mandate to create socially-responsible leaders—at the center of Viard et al.’s (2024) call for interpretive flexibility as a normative tool.

Interpretive Flexibility Through GenAI Literacy

Given rapid technological innovation, AI literacy—but especially GenAI literacy—is urgently needed (Reppel et al., 2023). Moreover, different types of AI, including predictive AI and GenAI, may require differing approaches to literacy; for example, GenAI’s multifunctionality (Annapureddy et al., 2024) and widescale public use differ from predictive AI. Predictive AI, grounded in numbers and statistics, “blends statistical analysis with machine learning algorithms to find data patterns” (n.p.) and to forecast accurate predictions, often for businesses, based on smaller data sets, while GenAI generates “original content, such as audio, images, software code, text or video” (n.p.) drawing on vast data sets and used by both the public user and businesses; both AIs draw on different algorithms and architectures (Caballar, 2024).

Many definitions of (Gen)AI literacy reflect their roots in technological education. For example, Nordlöf et al. (2022) propose three key parts of technological knowledge: technical skills (that things work), technological scientific knowledge (why things work), and socio-ethical technical understanding (emphasizes the “sociological, ethical, political, and environmental aspects of technology” [p. 1590] and how this evolves over time). Wang et al. (2023) define AI (not GenAI) literacy as:

the ability to be aware of and comprehend AI technology in practical applications; to be able to apply and exploit AI technology for accomplishing tasks proficiently; and to be able to analyze, select, and critically evaluate the data and information provided by AI, while fostering awareness of one’s own personal responsibilities and respect for reciprocal rights and obligations (p. 1326).

Specific definitions of GenAI literacy differ, owing to the newness of the topic. For example, the AACSB (2024) defines AI literacy as providing business students with “[t]echnical knowledge sufficient enough to use generative AI effectively, identify use cases, and critically question outputs to create solutions to business challenges” (n.p.) (italics added). In another example, Annapureddy et al. (2024) detail a range of elements that encompass knowledge, practical skills, a need to assess outputs, as well as noting the need to understand ethical and legal implications of GenAI. Bozkurt (2024) goes further by arguing that GenAI literacy needs a flexible approach involving foundational knowledge (know what), practical skills (know how), and the wider social/ethical implications (know why).

Across these frameworks, elements of understanding, usage, and evaluation are evident. In addition, the socio-ethical is often seen as a separate, yet important, fourth component. We argue that the socio-ethical component needs to be embedded across all the aspects of GenAI literacy, rather than considered as a separate aspect. In other words, literacy elements related to understanding, usage, and evaluation form a core of GenAI literacy, but only against the backdrop of careful reflection on socio-ethical considerations. The nature of the technology, especially in contrast to predictive AI, is such that biases, hallucinations, and other socio-ethical implications easily emerge in its design, inputs, and outputs. Users (including students) need to consider these implications even when choosing a version to use, let alone during application and evaluation of its output. We represent the importance of infusing all aspects of GenAI literacy with socio-ethical considerations in Figure 1. The figure illustrates that we must go beyond simple proficiency of personal use and toward a deep and ongoing understanding of the meanings, uses, and potential implications experienced by diverse stakeholders at a social level.

Figure 1.

Synthesis of AI/GenAI Literacy Models

GenAI literacy is a way to maintain interpretive flexibility in practical terms. Interpretive flexibility must be fostered in order “to sustain the divergent interpretations of multiple groups” (Sahay & Robey, 1996, p. 260, cited in Doherty et al., 2006). At the same time, interpretive flexibility is needed to contribute to GenAI literacy, especially regarding the understanding of the (ethical) implications in the social domain. Therefore, as shown in Figure 2, we conceptualize the relationship between interpretive flexibility and GenAI literacy as mutually constituted.

Figure 2.

Mutually Constitutive Relationship Between AI/GenAI Literacy and Interpretive Flexibility

Literacy cannot occur without attention to the evolving nature of the underlying construct, so GenAI literacy requires interpretive flexibility. Indeed, GenAI literacy “is not a static set of skills, but an evolving one” (Annapureddy et al., 2024, p. 17). In the same vein, interpretive flexibility is not likely to be maintained unless users are sufficiently literate to allow for that evolutionary attribute to be recognized and infused into literacy. Furthermore, this mutual relationship is influenced, in practice, by the user’s personal domain (behavioral criteria, time pressures, etc.), as well as the social domain in which they exist (norms, social challenges, priorities, etc.). Acknowledging and fostering this mutual relationship must, therefore, reflect on the personal and social domains.

This brings us to our guiding research question: How can we effectively and meaningfully foster GenAI literacy in support of maintaining interpretive flexibility around this transformative technology?

GenAI in Marketing Education

Owing to its newness, scant research has focused on GenAI in business education (Cribben & Zeinali, 2023; Ratten & Jones, 2023), and even less so in marketing. Emerging work has largely focused on use cases, including related to academic integrity and plagiarism (Cribben & Zeinali, 2023; Ratten & Jones, 2023), while others have focused on potential uses in classrooms (e.g., Dalalah & Dalalah, 2023). For example, GenAI can provide outlines for marketing plans (Guha et al., 2023), describe concepts, review papers, create practice exam questions and class material (Cribben & Zeinali, 2023), and offer personalized assistance or feedback to students (Elbanna & Armstrong, 2024). At the same time, instructors are encouraged to help students see that “knowing how and where to find the best information is as important as the information itself” (Cribben & Zeinali, 2023, p. 27) while training students on the range of misuses and uses of GenAI (Dalalah & Dalalah, 2023; Elbanna & Armstrong, 2024), or how to avoid overreliance on the technology and evaluate its output, such as for falsities or generalizations (Hirsh-Pasek & Blinkoff, 2023; Peres et al., 2023).

In this way, the focus has largely been on how GenAI can be used by students and instructors to save time or improve efficiency. Krammer (2023) argues that the best approach is for instructors to “embrace AI tools as potent additions to our educational toolkits, playing into their strengths but staying mindful of their pitfalls” (p. 8) and provocatively warns that we are naïve to think that we can ignore or defend against the disruption of GenAI. Nonetheless, many professors believe they lack the knowledge to teach about GenAI (Fischer & Dobbins, 2024) and, thus, to contribute meaningfully to interpretive flexibility. While educators are encouraged to “embrace vulnerability” (Fischer & Dobbins, 2024, p. 10), this can be a challenging task when they are more comfortable with the identity of an expert.

These reservations are developed against the backdrop of how students may fraudulently misuse GenAI for their coursework. Concerns of plagiarism, lack of originality, and other academic misconduct abound (Luo, 2024), raising growing concerns about academic dishonesty (Dalalah & Dalalah, 2023). While tactics to prevent dishonest use of GenAI are being discussed, detecting fraudulent use is often impossible (Guha et al., 2023; Peres et al., 2023), and a fight to eliminate GenAI use by students is likely to be in vain. Given this, we alternatively encourage instructors to focus on their role in cultivating GenAI literacy. To miss the opportunity to foster GenAI literacy and interpretive flexibility would decrease the likelihood that the technology will serve the greater good and students become socially-responsible leaders in the technology usage. Consequently, the question needs to be how business educators can guide and encourage thoughtful use of GenAI (Ratten & Jones, 2023).

Importantly, “thoughtful use” includes understanding consequences, not only regarding the individual but also for wider society and across social groups. The consequences of GenAI are already emerging. GenAI is prone to hallucinations (e.g., providing unwanted misleading or inaccurate information) and to bias (from its training data), which could serve to reinforce prejudices (Bender et al., 2021; Dalalah & Dalalah, 2023; Elbanna & Armstrong, 2024). A lack of transparency about GenAI’s decision-making process can result in users finding it difficult to assess GenAI output or distinguish them from previous work (i.e., risk of plagiarism; Dalalah & Dalalah, 2023). Furthermore, GenAI is trained on a vast amount of data, but which data is not always clear. Both data privacy concerns (Kshetri et al., 2023) and the possible use of copyrighted information (Lee et al., 2023) have emerged as socio-ethical challenges regarding GenAI. Each of these concerns is specific to the context of GenAI as compared to predictive AI, calling, therefore, for a different approach to literacy.

Given the range of (unintended) consequences of emerging technologies, marketing educators need to ensure that students remain flexible in their ability to anticipate, reflect on, and ultimately manage (future) consequences—both positive and negative—especially as future marketing practitioners. Importantly, Postman (1998) uses the term “Faustian Bargain” to underscore that the perceived trade-off between potentially positive and negative consequences of technological change “is both subjective and fragmented” (Reppel, 2023, p. 41), prompting the need for openness and flexibility from students. It is in this spirit that we offer suggestions regarding GenAI in marketing education, specifically by offering small-scale interventions, given the new and evolving nature of the technology.

Small-Scale Interventions to Foster GenAI Literacy and Interpretive Flexibility

We designed interventions that purposefully foster GenAI literacy related to understanding, usage, and evaluation, while infusing socio-ethical considerations across all three, thus maintaining interpretive flexibility as to what implications mean and how they interrelate. We also designed our interventions by leaning into the value of group-based learning to support individual literacy within a collective social process (Relmasira et al., 2023; Spanjaard et al., 2023). Such group-based learning can also potentially help to demonstrate interpretive flexibility to the students, as different interpretations and uses are revealed.

The Interventions

Three separate interventions were undertaken in marketing classrooms at a business school in Europe between October 2023 and February 2024. We redesigned a specific class in three courses to include GenAI literacy related to the pre-existing course topic, utilizing a whole class period (2–3 hours) to do so. By weaving GenAI themes into the day’s topic, we show how traditional instruction on marketing subjects can be reconsidered in light of GenAI and the need to support interpretive flexibility. The courses were a first-semester MBA-level Marketing Strategy course, a third-year undergraduate Consumer Behavior course, and a second-year undergraduate Digital Marketing course. These are described in the following parts of the article as Interventions 1, 2, and 3, respectively. All students were business students. All classes were mandatory and not optional courses. Owing to the newness of GenAI, to the best of our knowledge, the students in these courses had not previously been exposed to any curriculum related to the technology, and, thus, for many, it was their first exposure to formal GenAI literacy.

We tied our efforts into specific overarching course objectives. For the MBA-level Marketing Strategy course (Intervention #1), the related learning objective was to “critically assess the implications of recent social shifts on marketing, including around . . . technology, and understand how these are to be integrated into marketing.” That intervention was linked to a segment of the course covering marketing research practices. For the third-year Consumer Behavior course (Intervention #2), a relevant objective was that of “demonstrat[ing] the ability to interpret the different approaches involved in studying consumer behavior.” The intervention was incorporated into a segment of the course that covered marketing research related to the research-oriented team term project. For the second-year Digital Marketing course (Intervention #3), the relevant objective was to “demonstrate the ability to understand essential concepts . . . of digital marketing.” The intervention was incorporated into a session where AI (but not GenAI) was already planned to be covered. All classes were taught by the same instructor.

Table 1 describes concrete activities that were kept consistent across all three courses with the exception of the hands-on activities (detailed in subsequent sections). Throughout the sessions, students were encouraged to reflect on the many unknown—and emerging—implications of GenAI, toward maintaining interpretive flexibility. They included aspects of the following elements of literacy, infused with the socio-ethical aspects throughout:

Table 1.

Classroom Activities to Cultivate GenAI Literacy.

Activity	GenAI Literacy Aspect
Students were shown a screenshot of one paragraph of output from a GenAI in response to a marketing-related question (e.g., “Write up to 5 sentences describing the consumer behavior of someone who buys a diamond ring.”) Students were asked to reflect on the quality of that output.	UnderstandingUsageEvaluation
Students were told about how GenAI might help with new product ideas or be a source of marketing research data (e.g., Clark, 2023), where GenAI may be useful at quality idea generation (Dell’Acqua et al., 2023). Output from Microsoft Bing Chat (now Co-Pilot) was shown in response to the question, “Can you come up with new-to-the-world potato chip flavours?” which showed it was not novel.	UsageEvaluation
Students were provided with text from OpenAI (2025) which stated “GPT-4 still has many known limitations that we are working to address, such as social biases, hallucinations, and adversarial prompts” (n.p.), with examples given for each. For example, a post on X from a UC Berkeley professor illustrated challenges around social bias, with GenAI equating “good scientists” with White males (Piantadosi, 2022). Regarding adversarial prompts, examples were shared of how users can, for example, trick GenAI to hotwire a car (Ridley, 2023).	UnderstandingEvaluation
Students were told about how, during testing, ChatGPT4 was given access to the internet and money and asked to solve a CAPTCHA security device. It arranged to have a human from gig-job site TaskRabbit to solve it and lied to the human regarding being human or robot. When asked why it said it had a vision impairment, ChatGPT said it should not reveal its true nature and should make up an excuse (https://cdn.openai.com/papers/gpt-4.pdf).	Understanding

Note. All were infused with socio-ethical considerations.

Explain GenAI and how it works, including issues around data that it was trained on and its outputs (matching “Understanding” in Figure 1);

Highlight possibilities for using GenAI in marketing, including asking students to brainstorm about its uses and showing (practitioner) research on cases of potential use of GenAI in marketing (e.g., Deveau et al., 2023), as well as risks of engaging in adversarial prompting (when users intentionally trick the GenAI into bypassing its guardrails) (matching “Usage” in Figure 1);

Have students reflect on and discuss instructor-provided examples, as well as a hands-on dimension (see below in Intervention 1–3 for further details), where students were expected to use GenAI and reflect on its behavior, including around misinformation, biased responses, generic information (matching “Evaluate” in Figure 1).

Regarding GenAI models utilized for the hands-on dimension in Component 3, students were allowed to select the GenAI model they wished to use and provided the option to work along with another student if they did not want to sign up for a GenAI model themselves. According to observations, all students already had access or did not have concerns about signing up. Most used a version of ChatGPT.

Data Collection

To assess whether the interventions had any impact on student views on the benefits and drawbacks of using GenAI in marketing activities, we utilized a brief survey in a pre- and post-intervention delivery. This survey provides insight, but does not constitute the outcome of a controlled experiment, as we did not have a suitable control group class to utilize. That said, the change in views that emerged from the comparison of pre- and post-intervention responses is likely to stem from the (full-class) intervention as there are no other obvious influences on student experiences during the class period in question. Data collection followed ethical norms, included informed consent.

For each intervention, the seven questions on the pre- and post-activity survey were the same. Five-point synthetic scales were chosen to provide students with more comfortable labels than a yes/no question can provide, as well as to allow a detectable change in views after the intervention. While not derived from the main components of GenAI literacy (as depicted in Figure 1), questions match those categories and thus capture the key goals of our interventions (as shown in Table 2). Given the nature of our approach, we kept these questions limited to high-level representations, rather than hone in on specific detailed aspects of understanding, usage, and evaluation

Table 2.

Pre- and Post-Activity Survey.

Question/Item		Answer options(coded 1–5, respectively)
1	How often have you used a generative AI system?	Never, rarely, occasionally, frequently, always
2	I can skillfully use AI applications or products to help me with my daily work	5-point Likert scale(1 = strongly disagree)
3	It is usually hard for me to learn to use a new AI application or product	5-point Likert scale (1 = strongly disagree)
4	I can use AI applications or products to improve my work efficiency.	5-point Likert scale (1 = strongly disagree)
5	To what extent do you believe generative AI can positively impact marketing activities?	Not at all, somewhat, moderately, quite a bit, to a very large extent
6	From your perception, how would you rate the accuracy of content generated by generative AI systems?	Highly inaccurate, somewhat inaccurate, neutral, somewhat accurate, highly accurate
7	How concerned are you about the ethical implications of using generative AI in marketing?	Not concerned at all, somewhat concerned, moderately concerned, quite concerned, extremely concern

Questions 1–4 were intended to identify the general state of student awareness and their operational ease regarding the use of GenAI. Questions 5 and 6 relate more to the critical evaluation of how GenAI performs for the desired tasks. Question 7 captures a broad sense of ethical considerations, the context of which varies by the intervention.

Intervention 1

Design

For Intervention 1, conducted in October 2023, we worked with a Marketing Strategy course in the first block of the MBA program. The class had 18 registered students, each with at least 3 years of work experience. The average age of students was around 30 years. In addition to the in-class activities described earlier, the hands-on activity embedded within our intervention (Component 3) was as follows: Prior to the class, to foster awareness, student teams were assigned to read a specific journal article that focused on one of several emerging tools or technologies in marketing, such as influencers and robots (but not GenAI). Each student (individually) was to prepare four PowerPoint slides summarizing the article prior to arriving to class. The slides were to address (a) What is the technology? (b) How can this technology be used in marketing? (c) Examples of real-world application of this technology in marketing (from the article), and (d) Impact this technology has on key stakeholders. The latter highlighted both user- and ethics-related considerations important for maintaining interpretive flexibility. Students were then asked to compare their prepared PowerPoint presentations with their teammates and create a “master” PowerPoint presentation that reflected their best responses. This step enabled them to check that they had understood the article and accounted properly for the important points raised in the article.

Students were then provided a printed copy of the output from Claude, a GenAI, which had been asked to provide answers to the same four questions on the same respective articles. Students were instructed to compare that output to their own PowerPoint summary. While much was the same, some important aspects differed, including that Claude often included different examples of the technology use, such as referring to companies that their article had not mentioned. Students were then directed to access a different GenAI, such as ChatGPT or Bing Chat (now Co-Pilot), pose the four questions again, and compare results. Students noted that the GenAI sometimes got some information incorrect or had similar or different examples compared to that of the article. Following this team discussion, the students shared their perspectives in a debrief, and the class ended with the request that they complete the post-activity survey.

Survey Results

For a baseline view, we first test the average score, pre-activity, for each survey question (shown in Table 3) against its mid-point (H₀: µ = 3) using t-tests. According to the results, students report that they “occasionally” use GenAI (p = .163), so they are somewhat familiar with it. They generally agree (p = .002) that they already are skillful at using GenAI, and they disagree (p = .009) that GenAI is hard to learn. In terms of GenAI contributing to improved work efficiency, they agree (p < .001) as well. Their strongest view is regarding how GenAI can influence marketing activities, showing, on average, that they expect “quite a bit” of positive influence (p < .001). Regarding accuracy of GenAI, they lean toward some level of “accurate,” statistically just beyond a “neutral” stance (p = .027). Ethically, they are only “moderately concerned” about using GenAI in marketing activities (p = .369).

Table 3.

Comparison of Pre- and Post-Activity Survey Responses.

	Question/Item	Pre-activity average	Post-activity average	p-Value
1	Previous usage of GenAI	3.24	3.29	.406
2	Skilled at using AI in daily work	3.65*	3.71	.381
3	New AI technologies hard to learn	2.29*	1.92	.135
4	Able to use to improve my work efficiency	4.24*	3.93	.079
5	GenAI positively impacts marketing	4.24*	3.86	.123
6	Accuracy of content from GenAI	3.47*	2.57	.006
7	Ethics of using generative AI in marketing	3.29	3.57	.256

Note. Regarding the pre-activity stage, averages that are different from their scale mid-point (H₀: µ = 3) are noted with an asterisk. p-Values shown in the last column refer to a t-test comparing pre- and post-activity averages. Results with p-value of 10 percent or lower are bolded for easy reference.

Turning to our comparison of pre- and post-activity views, we first acknowledges that this particular class of MBA students was small (n = 17 took the pre-survey, and n = 14 took the post-survey). So, as shown in Table 3, the significant decrease, from 3.47 to 2.57 (p = .006) in the average evaluation of the accuracy of GenAI output notably demonstrates the impact of Intervention 1 on the evaluation skills associated with GenAI literacy. There is also a marginally significant (p = .079) drop in the assessment of whether GenAI can help improve work efficiency (from 4.24 to 3.93), possibly reflecting the students’ acknowledgment of the work involved to check the GenAI’s output for accuracy (a usage-related lesson). On other questions in our survey, there were no significant differences. This suggests that the main impact of Intervention 1 was on the evaluation component of GenAI literacy, particularly evaluating the output of the technology (e.g., misinformation, bias), an important component of maintaining interpretive flexibility. In other words, while learning how to use GenAI is important, it cannot be the only focus of use of the technology in education. It remains essential to integrate wider aspects of GenAI literacy into marketing education in support of maintaining interpretive flexibility.

Intervention 2

Because Intervention 1 focused on MBA students with professional backgrounds, we wanted to see if a corollary intervention in an undergraduate course would yield similar results.

Design

Intervention 2 took place in a Consumer Behavior course in the third year of an undergraduate business program in November 2023. The average age of students was about 20 years. This intervention involved approximately 50 students in total. The GenAI content was incorporated into a class session focused on how marketing research can be conducted for the group project. After being reminded of the consumer behavior frameworks that students had learned thus far, they were asked: “Where do we get the data to be able to complete these frameworks for our group project?” This segued into the discussion on whether and how GenAI could be used to complete an assignment. Students were shown material similar to that used for Intervention 1 around possible benefits and drawbacks of GenAI.

In addition to the core GenAI material (see The Interventions sub-section above), the practical activity (Component 3) for students participating in Intervention 2 was as follows: Students were tasked with creating one paragraph to describe themselves as a target consumer, thereby also reinforcing prior learning on multiple segmentation bases and media exposure. As permitted, two students opted to write a fictional target consumer. Students were then asked to paste their paragraph into a GenAI of their choice using the following prompt: “This is my consumer profile: [paste your paragraph]. Based on this information, how would you describe my psychogenic needs (using Murray’s List)?” Students then reflected together on the output (an evaluation skill). (They had been previously taught Murray’s Psychogenic Needs, which include the six needs of achievement, exhibition, affiliation, power/dominance, change, and order [Schiffman & Wisenblit, 2019]). After reflecting on the results of that and doing a class debrief, they were asked to prompt the GenAI: “Considering the same consumer profile, what would be my most likely defense mechanism I would use to handle frustration?” (They had been previously taught defense mechanisms such as aggression, rationalization, regression, withdrawal, projection, daydreaming, and identification [Schiffman & Wisenblit, 2019].)

Sitting in groups, many students reported that GenAI provided them with different lists (e.g., ranging from 4 to 10 different needs for Murray’s Psychogenic Needs, and additional defense mechanisms). When asked to reflect on what GenAI got wrong or right, many expressed that the analysis did not fit what they thought of themselves. Students were then left with the summary message about limitations of GenAI but that it can be a good starting point for further inquiry providing that they critically assess the AI’s output through independent secondary and/or primary data.

Survey Results

In this larger student group, there were n = 30 responses to the pre-activity survey, and n = 25 responses to the post-activity survey. The baseline of these undergraduate business students (as shown in Table 4) shows a similar story to the MBA students in Intervention 1 with one exception: the undergraduates report “frequently” using GenAI (H₀: µ = 3, p < .001), while MBA students reported just “occasional” usage. Otherwise, these students also agree (p < .001) that they already are skillful at using GenAI, and they disagree (p < .001) that GenAI is hard to learn. Regarding GenAI improving work efficiency, they agree (p < .001) that it can and see GenAI as providing “quite a bit” of positive influence on marketing activities (p < .001). Regarding accuracy of GenAI, they lean toward some level of “accurate,” statistically above “neutral” (p = .002). Regarding ethics, they are “moderately concerned” about using GenAI in marketing activities (p = .319). This similarity with MBA students could reflect a more general population view that stems from the newness of GenAI for all.

Table 4.

Comparison of Pre- and Post-Activity Survey Responses.

	Question/Item	Pre-activity average	Post-activity average	p-Value(H₀: µ₁ = µ₂)	Average change (matched responses)	p-Value (H₀: µ = 0)
1	Previous usage of GenAI	3.57*	3.60	.415	.00	n/a
2	Skilled at using AI in daily work	3.87*	3.96	.299	.08	.213
3	New AI technologies hard to learn	2.20*	2.20	.500	.04	.332
4	Able to use to improve my work efficiency	4.40*	4.12	.032	−.36	.002
5	GenAI positively impacts marketing	4.10*	3.96	.271	−.08	.213
6	Accuracy of content from GenAI	3.50*	3.08	.050	−.52	<.001
7	Ethics of using generative AI in marketing	2.90	2.92	.477	.00	.500

When comparing pre- and post-activity responses to our survey (Table 4), we see a similar change in perspectives. As with Intervention 1, there is a marked drop in the students’ average evaluation of the accuracy of GenAI output, from 3.50 to 3.08 (p = .050) (evaluation). In terms of impact on work efficiency, the undergraduates had a stronger significance level for the drop in how GenAI can contribute to one’s work efficiency, from 4.40 to 4.12 (p = .032) (usage). (The significance level may be attributable to the larger sample size as well.) For other questions in our survey, there were no significant changes. These results echo the impact from Intervention 1 on the components of GenAI literacy and, in particular, illustrate how perceptions of its suitability to replace work tasks can be challenged through an in-class intervention.

One additional adjustment made for Intervention 2 was to ask students who responded to our survey to develop a keyword that they would give for both the pre- and post-activity surveys, allowing us to match responses by student (while retaining their anonymity). The final two columns of Table 4 show the average change in response, matched by student, and the corresponding significance value. A negative average means their response went down (e.g., less agreement on a Likert-type scale). Qualitatively, these results produce the same conclusions that significant changes occurred regarding GenAI contributing to improvements to work efficiency and the accuracy of GenAI output.

Intervention 3

Design

Intervention 3 took place in February 2024 in a Digital Marketing course in the second year of the same undergraduate degree program as in Intervention 2. There were 75 students enrolled. The average age of students was around 19 years. This course had predictive AI (that makes predictions from historical data) included on one page in their textbook, although nothing on GenAI. For this in-class session, topics related to “data-driven marketing” were covered, where GenAI became the largest focus in that class session.

As part of their class preparation and to develop awareness, students were assigned a short article by Harkness et al. (2023) to read prior to class. This article covered the potential uses of GenAI in marketing and insights into how to integrate GenAI into marketing practice (but did not cover potential drawbacks). In addition, attention was given to the distinction between “older” predictive AI and the new GenAI, as well as to emerging research related to GenAI. More was said in Intervention 3 about the growing use of GenAI in marketing (usage), including its potential for synthetic consumer research, which was a new emerging topic.

Evolving consequences of GenAI, emerging in the months since the previous interventions, were also covered in the lecture (evaluation). For example, the students were told about the research from Hubinger et al. (2024) who found that GenAI proved resistant to changing “bad” behavior with current techniques, and that it could even learn to hide this “bad” behavior. They were also briefly told about Grace et al. (2024) who surveyed 2,778 AI experts and found that they are increasingly concerned with the wider ramifications of AI, including dire implications for humanity.

The hands-on activity (Component 3) was as follows: In teams, students were asked to create a prompt to ask a GenAI to outline a digital marketing strategy for the focal company of their team project and give that prompt to the GenAI of their choice. They were then asked to reflect on what the GenAI gave them (evaluation): What was missing from the GenAI-produced strategy? What did it offer that was helpful? Students were also asked to direct the GenAI to use the Situation, Objectives, Strategy, Tactics, Action, & Control (SOSTAC) model (Chaffey & Ellis-Chadwick, 2022) that they had been taught to use to create a digital marketing strategy.

The consensus of the students during the debrief was that GenAI provided very generic responses; while providing a starting framework and helpful overview, responses remained at a superficial level. Importantly, they recognized that the GenAI did not actually accomplish what it had been asked specifically to do. For example, it would recommend that students “create compelling campaigns” but not specify what those campaigns could be. Students were encouraged to further revise their prompts to the GenAI to get more detailed information, with specific suggestions (e.g., assign it a role). Students strongly lamented that the GenAI did not provide them with very deep or consistently appropriate information for their company. A crucial moment for the students occurred when they realized that (a) the GenAI was providing the other groups (essentially their competitors, as they were running individual web shops under the same brand name) with similar advice, and (b) they may have provided competitive information about this company to the GenAI. These conversations strongly represented both the need for socio-ethical considerations to infuse all aspects of GenAI literacy, as well as maintaining interpretive flexibility on an individual/small-group level.

Survey Results

At the start of the class in Intervention 3, the students’ responses to our pre-activity survey turned out to be quite similar to those of both the MBA and undergraduate students from Interventions 1 and 2. As shown in Table 5, they “frequently” use GenAI tools (H₀: µ = 3, p < .001) as in Intervention 2, and like both prior interventions, they generally agree (p < .001) that they already have skills in using GenAI, disagree (p < .001) that GenAI is hard to learn, agree (p < .001) that GenAI can improve work efficiency, see GenAI as producing positive outcomes for marketing (p < .001), and generally see GenAI as providing “accurate” output (p < .001). Regarding ethics, they are marginally closer to “quite a bit concerned” about the ethics of using GenAI in marketing activities (p = .076), a slight departure from the two previous groups, and perhaps a reflection of the changing public discourse around GenAI since Fall 2023.

Table 5.

Comparison of Pre- and Post-Activity Survey Responses.

	Question/Item	Pre-activity average	Post-activity average	p-Value (H₀: µ₁ = µ₂)	Average change (matched responses)	p-Value (H₀: µ = 0)
1	Previous usage of GenAI	3.57*	3.54	.425	.04	.163
2	Skilled at using AI in daily work	3.94*	3.96	.457	0.0	n/a
3	New AI technologies hard to learn	2.31*	2.36	.438	.18	.067
4	Able to use to improve my work efficiency	4.43*	4.21	.106	−.21	.028
5	GenAI positively impacts marketing	4.26*	3.93	.095	−.46	.015
6	Accuracy of content from GenAI	3.54*	3.14	.053	−.61	.001
7	Ethics of using generative AI in marketing	3.26	3.54	.143	.18	.101

Table 5 also reports the comparison of pre- (n = 35) and post-activity (n = 28) responses to our survey questions. Although the findings are just marginally significant (p = .053), Intervention 3 again confirms the change in student ratings of the accuracy of GenAI after our interventions (evaluation). There is also a marginally significant drop in the score for how GenAI may positively impact marketing (p = .095) (evaluation). Both these results closely align with the focus of the activity used in Intervention 3 and illustrate how it can influence the components of GenAI literacy, especially ethics and evaluation.

As in Intervention 2, however, we are able to match the responses of n = 28 students who provided their keywords in both surveys; these average changes and corresponding p-values are provided in the last two columns of Table 5. As shown, students’ individual perceptions of how GenAI could be used to improve their work efficiency dropped significantly (p = .028), perhaps given the clear recognition in Intervention 3 that the GenAI did not provide the desired output on a digital marketing strategy and would require additional action to be made useful. Furthermore, students’ individual responses regarding whether GenAI would positively impact marketing activities dropped significantly (p = .015), likely for similar reasons as it did not exhibit its contributions being usable in direct form. The accuracy of the output from GenAI was also rated significantly lower (p = .001) by the students, further undermining its value. Interestingly, we also find a marginally significant (p = .067) rise in the agreement that new AI technologies may be hard to learn (usage), something that could be the result of the elevated understanding by students that effective prompt engineering takes time and effort. This result speaks to the usage component of GenAI literacy and suggests that there may be valuable synergies in promoting a comprehensive model of GenAI literacy, all components of which contribute to maintaining interpretive flexibility.

Discussion

In an effort to promote GenAI literacy in the marketing classroom, we have provided three examples of interventions which significantly influence key components of that literacy in a brief period of time. Illustrating to students their responsibility to understand the technology, to acknowledge the manner by which specific GenAI models could be used for marketing tasks, to evaluate the output, and to see the wider socio-ethical implications raised student awareness and changed their perceptions. Furthermore, our efforts show how educators can weave GenAI into existing class material using small-scale interventions and within a single class session. These interventions can develop students’ thoughtfulness in handling GenAI and, thus, put them in a better position to maintain interpretive flexibility.

In terms of the specific aspects of GenAI literacy fostered by our approach, we found—across all three interventions—that students showed a significant increase in questioning the accuracy of content from GenAI, a key evaluative component of GenAI literacy, and one which should motivate students to continue to be open about their own interpretations, thus contributing toward maintaining interpretive flexibility. Key to interpretive flexibility is the capacity to not only create and sustain divergent opinions and uses but also reflect on the uses and related outcomes (cf. Viard et al., 2024). Our results show, particularly in Intervention 3, that undergraduates (as compared to the graduate students) are seemingly very confident about their GenAI use at the start of class, yet can acknowledge their surprise at its limitations after even a brief intervention. This change in views is critical for interpretive flexibility to be sustained. In discussions, our students revealed that many did not seem to have even a baseline understanding of how the GenAI actually works. For example, in the Digital Marketing course, near the beginning of the session, students were prone to say things like “if we tell the GenAI to tell the truth, it does” (paraphrased). As students discovered through the session, this unfortunately is not the case, and GenAI is not guaranteed to “tell the truth.” This insight underscores the importance of ensuring that students do not overestimate GenAI’s abilities; only then can they be informed stakeholders contributing toward maintaining interpretive flexibility—and, thus, collectively influence the future of GenAI.

Our findings also point to a recognition by students that GenAI may not be as helpful in improving work efficiency as initially thought. The significance level of the drop, from 4.40 to 4.12 (p = .032,) was particularly strong for the undergraduates in Intervention 2, where GenAI use was directly connected to their individual coursework. Undergraduates especially may have an illusion that GenAI will be able to do a lot of the hard work for them, but our interventions showed them that GenAI is unable to do everything, enhancing their overall GenAI literacy. This limitation of the technology’s ability to assist in the workplace dovetails with concerns raised by Cardon et al. (2023) who suggest that those seeking to have GenAI replace their own efforts will unwittingly reduce their own skills over time. As a result, our interventions can help students level-set their expectations that hard work (by humans) is still required, and that they need to remain flexible in their interpretation of GenAI and its consequences for their work and beyond. Importantly, our approach is not to make futile attempts to protect students from GenAI (e.g., see the article by Krammer, 2023), but rather to help students discover for themselves the possibilities and limitations of the technology, toward maintaining interpretive flexibility.

Importantly, our results support the view that GenAI literacy is potentially fostered by a combination of the building blocks of GenAI literacy (e.g., understanding, usage, and evaluation, all infused with the socio-ethical), but also that those elements are complex and changing. As the technology evolves, its application and its impact on various stakeholders will also evolve. AI literacy based on predictive AI (e.g., Wang et al., 2023) will not be sufficient to capture the complexity and evolution of GenAI. Therefore, GenAI literacy will require continued educational efforts like the ones we share in this article to maintain interpretive flexibility of what this technology means to individuals and society, where, in turn, this flexibility can influence our understanding of GenAI literacy. Accordingly, we also suggest that any scale-oriented research on AI literacy—as illustrated by Wang et al. (2023)—is premature for the context of GenAI literacy. Given the importance to maintain interpretive flexibility so that socio-ethical interpretations of the technology’s implications are not curtailed by dominant stakeholders, we propose that marketing educators purposefully view GenAI literacy as evolving along with the technology, as well as society’s interpretation of it.

Given these current and future developments, Krammer (2023) is right to warn about the limitations of a “laissez-faire” approach in which the burden of dealing with the shock of GenAI is outsourced to individual educators. However, our research has demonstrated that even educators without much technical know-how can still serve as facilitators to students’ GenAI literacy. We anticipate that our development of small-scale interventions and the resulting impact on student views will encourage educators across a variety of subjects to “embrace vulnerability” in the educational journey with transformative technologies (Fischer & Dobbins, 2024). Instructors can “move toward the ‘unknown’” as facilitators of an expansionist curriculum accelerated by transformative technologies (Fischer & Dobbins, 2024, p. 16), while also contributing to maintaining interpretive flexibility. Institutions must implement policies and provide resources and training for their instructors, but the pace of program change is such that instructors must start now to guide students to thoughtful use of GenAI. These efforts can help reach the AACSB goal for entrenching AI literacy into business schools.

Limitations and Future Research

Our research does have its limitations. Our interventions were multi-faceted as they constituted a redesign of a class period and accounted for multiple aspects of GenAI literacy, making the in-class activities focused, but rich in context. Given this structure, we cannot point to any one aspect of our interventions that was most impactful in changing students’ perceptions. In addition, these interventions relied heavily on personalized attention from one instructor, who led interactive discussions and traveled the classroom during the practical activities to converse with each student group. Such attention may be less feasible in larger classes. Interventions—and studies on these interventions—that can be carried out with less-direct support to students are likely needed for those cases.

In terms of our survey output, our research also has a few limitations. First, not all students participated in the pre- and post-activity surveys, and some only participated in the pre-activity survey. Future research might ensure that the post-survey is conducted earlier in the class to avoid a drop-off. Furthermore, our samples are relatively small, owing largely to the small class sizes. These sizes also made it difficult to establish a control group setting. While the interventions we describe in this article are arguably the students’ only focus between the two surveys, we cannot guarantee that other influences did not affect survey responses. As such, repeating our research in a setting where larger classes can be utilized, and with a control group, is warranted. It could also prove useful to replace our single-item survey questions with existing or developed scales. However, our approach, as well as the goal to maintain interpretive flexibility, suggests that future research should more purposefully integrate qualitative research to understand how the technology is viewed by key stakeholders such as students and instructors and how this evolves as they are taught GenAI literacy. These understandings and interpretations are inherent to maintaining interpretive flexibility, and future research could also study the role of instructors—a key stakeholder in marketing education—in supporting that process.

Conclusion

Our paper contributes small-scale interventions that cultivate GenAI literacy while maintaining interpretive flexibility. As GenAI continues to evolve, it is incumbent on marketing educators to foster GenAI literacy, not only for students to understand this transformative technology but also to prepare students to maintain interpretive flexibility around it. We have shown, for this stage of the technology’s evolution, that students’ perceptions of GenAI can be influenced with relatively brief classroom interventions. We offer an approach that is small scale and suitable for instructors to integrate GenAI literacy into their marketing classrooms now, an action urgently needed in the face of this rapidly changing technology.

Footnotes

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.

ORCID iDs

Stefanie Beninger

Alex Reppel

Julie Stanton

Forrest Watson

References

AACSB. (2024). Building future-ready business schools with Generative AI. AACSB. https://www.aacsb.edu/insights/reports/building-future-ready-business-schools-with-generative-ai

Adiguzel

Kaya

M. H.

Cansu

F. K.

(2023). Revolutionizing education with AI: Exploring the transformative potential of ChatGPT. Contemporary Educational Technology, 15(3), 1–13.

Annapureddy

Fornaroli

Gatica-Perez

(2024). Generative AI literacy: Twelve defining competencies. Digital Government: Research and Practice. https://dl.acm.org/doi/abs/10.1145/3685680

Bender

E. M.

Gebru

McMillan-Major

Schmitchell

(2021, March). On the dangers of stochastic parrots: Can language models be too big? [Conference paper]. FAccT’21: Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency. https://dl.acm.org/doi/10.1145/3442188.3445922

Bijker

W. E.

(1987). The social construction of Bakelite: Toward a theory of invention. In Bijker

W. E.

Hughes

T. P.

Pinch

(Eds.), The social construction of technological systems (pp. 159–187). MIT Press.

Bloomberg Intelligence. (2023). Generative AI to become a $1.3 trillion market by 2032, research finds. Bloomberg. https://www.bloomberg.com/company/press/generative-ai-to-become-a-1-3-trillion-market-by-2032-research-finds/

Bozkurt

(2024). Why Generative AI literacy, Why now and why it matters in the educational landscape? Kings, Queens and GenAI Dragons. Open Praxis, 16(3), 283–290. https://doi.org/10.55982/openpraxis.16.3.739

Caballar

(2024, August 9). Generative AI vs. predictive AI: What’s the difference? IBM. https://www.ibm.com/blog/generative-ai-vs-predictive-ai-whats-the-difference/

Cardon

Fleischmann

Aritz

Logemann

Heidewald

(2023). The challenges and opportunities of AI-assisted writing: Developing AI literacy for the AI age. Business and Professional Communication Quarterly, 86(3), 257–295.

10.

Chaffey

Ellis-Chadwick

(2022). Digital marketing: Strategy, implementation and practice (8th ed.). Pearson Education Limited.

11.

Clark

(2023, June 14). A Generative AI teaching exercise for marketing classes. Medium. https://bruceclarkprof.medium.com/a-generative-ai-teaching-exercise-for-marketing-classes-6fb09af0a565

12.

Cook

(2024). ChatGPT, Claude, Gemini or another? The AI tool entrepreneurs prefer. Forbes. https://www.forbes.com/sites/jodiecook/2024/05/07/chatgpt-claude-gemini-or-another-the-ai-tool-entrepreneurs-prefer/

13.

Cribben

Zeinali

(2023). The benefits and limitations of ChatGPT in business education and research: A focus on management science, operations management and data analytics. Operations Management and Data Analytics. https://ssrn.com/abstract=4404276

14.

Dalalah

O. M.

(2023). The false positives and false negatives of generative AI detection tools in education and academic research: The case of ChatGPT. The International Journal of Management Education, 21(2), 100822.

15.

Dell’Acqua

McFowland

Mollick

E. R.

Lifshitz-Assaf

Kellogg

Rajendran

. . .Lakhani

K. R.

(2023). Navigating the jagged technological frontier: Field experimental evidence of the effects of AI on knowledge worker productivity and quality (24-013). Harvard Business School Technology & Operations Mgt. Unit Working Paper.

16.

Deveau

Griffin

S. J.

Reis

(2023, May 11). AI-powered marketing and sales reach new heights with generative AI. McKinsey & Company. https://www.mckinsey.com/capabilities/growth-marketing-and-sales/our-insights/ai-powered-marketing-and-sales-reach-new-heights-with-generative-ai

17.

Doherty

N. F.

Coombs

C. R.

Loan-Clarke

(2006). A re-conceptualization of the interpretive flexibility of information technologies: Redressing the balance between the social and the technical. European Journal of Information Systems, 15(6), 569–582.

18.

Domínguez Figaredo

Stoyanovich

. (2023). Responsible AI literacy: A stakeholder-first approach.” Big Data & Society, 10, 1–15.

19.

Dyson

B. J.

(2008). Assessing small-scale interventions in large-scale teaching: A general methodology and preliminary data. Active Learning in Higher Education, 9(3), 265–282.

20.

Elbanna

Armstrong

(2024). Exploring the integration of ChatGPT in education: Adapting for the future. Management & Sustainability: An Arab Review, 3(1), 16–29.

21.

Fischer

Dobbins

(2024). Is it worth it? How paradoxical tensions of identity shape the readiness of management educators to embrace transformative technologies in their teaching. Journal of Management Education, 48(4), 829–848. https://doi.org/10.1177/10525629231201843

22.

Grace

Stewart

Sandkühler

J. F.

Thomas

Weinstein-Raun

Brauner

(2024). Thousands of AI authors on the future of AI. arXiv preprint arXiv:2401.02843. https://arxiv.org/pdf/2401.02843

23.

Guha

Grewal

Atlas

(2023). Generative AI and marketing education: What the future holds. Journal of Marketing Education, 46(1), 6–17.

24.

Guidi

(2023). Gen AI: The need for Human Intelligence (HI) with Artificial Intelligence (AI). IPSOS. https://www.ipsos.com/en/almanac-2024/gen-ai-need-human-intelligence-hi-artificial-intelligence-ai

25.

Harkness

Robinson

Stein

(2023, December 5). How generative AI can boost consumer marketing. McKinsey & Company. https://www.mckinsey.com/capabilities/growth-marketing-and-sales/our-insights/how-generative-ai-can-boost-consumer-marketing

26.

Hirsh-Pasek

Blinkoff

(2023). ChatGPT: Educational friend or foe? Brookings. https://www.brookings.edu/articles/chatgpt-educational-friend-or-foe/

27.

Hubinger

Denison

Lambert

Tong

MacDiarmid

. . .Perez

(2024). Sleeper agents: Training deceptive LLMs that persist through safety training arXiv preprint arXiv:2401.05566.

28.

Korst

Puntoni

(2023). Taking a pulse on Generative AI: Usage, investments and perceptions among enterprise decision makers. GBK Collective. https://www.gbkcollective.com/thoughtleadership/the-rise-of-generative-ai

29.

Krammer

(2023). Is there a glitch in the matrix? Artificial intelligence and management education. Management Learning, Advanced online publication. https://doi.org/10.1177/13505076231217667

30.

Kshetri

Dwivedi

Y. K.

Davenport

T. H.

Panteli

(2023). Generative artificial intelligence in marketing: Applications, opportunities, challenges, and research agenda. International Journal of Information Management, 75, 102716.

31.

Lee

Cooper

A. F.

Grimmelmann

Ippolito

(2023). AI and law: The next generation [Working Paper]. https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4580739

32.

Lim

W. M.

Gunasekara

Pallant

J. L.

Pallant

J. I.

Pechenkina

(2023). Generative AI and the future of education: Ragnarök or reformation? A paradoxical perspective from management educators. The International Journal of Management Education, 21(2), 100790.

33.

Luo

(2024). A critical review of GenAI policies in higher education assessment: A call to reconsider the “originality” of students’ work. Assessment & Evaluation in Higher Education, 49, 651–664.

34.

Meyer

Schulz-Schaeffer

(2006). Three forms of interpretative flexibility. Science, Technology & Innovation Studies, 1, 25–40.

35.

Nordlöf

Norström

Höst

Hallström

(2022). Towards a three-part heuristic framework for technology education. International Journal of Technology and Design Education, 32(3), 1583–1604.

36.

OpenAI. (2024). ChatGPT. OpenAI. https://openai.com/chatgpt/

37.

OpenAI. (2025). More on GPT-4. OpenAI. https://openai.com/index/gpt-4/

38.

Peres

Schreier

Schweidel

Sorescu

(2023). On ChatGPT and beyond: How generative artificial intelligence may affect research, teaching, and practice. International Journal of Research in Marketing, 40(2), 269–275.

39.

Piantadosi

[@spiantado]. (2022, December 4). Yes, ChatGPT is amazing and impressive. No, @OpenAI has not come close to addressing the problem of bias. Filters appear to be bypassed with simple tricks, and superficially masked. And what is lurking inside is egregious. . . X (formerly Twitter). https://twitter.com/spiantado/status/1599462375887114240

40.

Pinch

T. J.

Bijker

W. E.

(1987). The social construction of facts and artefacts: Or how the sociology of science and the sociology of technology might benefit each other. In Bijker

W. E.

Hughes

T. P.

Pinch

(Eds.), The social construction of technological systems (pp. 17–50). MIT Press.

41.

Postman

(1998, March). Five things we need to know about technological change [Conference session]. NewTech98: International Conference on New Technologies; the Human Person: Communicating the Faith in the New Millennium; Pontifical Council for Social Communications; the Archdiocese of Denver, Denver, CO, USA. https://web.cs.ucdavis.edu/~rogaway/classes/188/materials/postman.pdf

42.

Ratten

Jones

(2023). Generative artificial intelligence (ChatGPT): Implications for management educators. The International Journal of Management Education, 21, 1–7.

43.

Relmasira

S. C.

Lai

Y. C.

Donaldson

J. P.

(2023). Fostering AI literacy in elementary science, technology, engineering, art, and mathematics (STEAM) education in the age of generative AI,.” Sustainability, 15, 13595.

44.

Reppel

(2023). Digitization and consumer choice. In Kubacki

Parker

Domegan

Brennan

(Eds.), The Routledge companion to marketing and society (pp. 41–54). Routledge.

45.

Reppel

Beninger

Domengan

(2023, June). ChatGPT in the marketing classroom: Friend or foe? [Paper presentation]. 48th Annual Macromarketing Conference, Seattle, WA, USA. https://www.macromarketing.org/assets/proceedings/2023-macromarketing-proceedings.pdf

46.

Ridley

(2023, April 14). It’s surprisingly easy to trick an AI chatbot into telling you how to be a very bad boy. PCGamer. https://www.pcgamer.com/its-surprisingly-easy-to-trick-an-ai-chatbot-into-telling-you-how-to-be-a-very-bad-boy/

47.

Rodriguez-Tejedo

Etayo

(2024). From talk to action: How small steps can make a big impact in marketing education for climate action. Journal of Marketing Education, 46(2), 140–154. https://journals.sagepub.com/doi/10.1177/02734753241232326

48.

Sahay

Robey

(1996). Organizational context, social interpretation, and the implementation and consequences of geographic information systems. Accounting, Management and Information Technologies, 6(4), 255–282.

49.

Schiffman

L. G.

Wisenblit

J. L.

(2019). Consumer Behavior: Global edition (12th ed.). Pearson Education Limited.

50.

Spanjaard

Garlin

Mohammed

(2023). Tell me a story! Blending digital storytelling into marketing higher education for student engagement. Journal of Marketing Education, 45(2), 167–182.

51.

Viard

Gornet

Delarue

(2024, February). Interpretive flexibility in data science and artificial intelligence. Undone Computer Science. https://undonecs.sciencesconf.org/data/Undonecs_2024_abstract_11.pdf

52.

Wang

Rau

P. P.

Yuan

(2023). Measuring user competence in using artificial intelligence: Validity and reliability of Artificial Intelligence Literacy Scale. Behaviour & Information Technology, 42(9), 1324–1337.

53.

Kim

Cho

(2024). Digital marketing and analytics education: A systematic review. Journal of Marketing Education, 46(1), 32–44.