ChatGPT, AI Advertising, and Advertising Research and Education

Since ChatGPT was released on November 30, 2022, followed by Microsoft’s announcement of its artificial intelligence (AI)–powered new Bing search engine on February 7, 2023, and Google’s Bard release on March 21, 2023, it seems AI generally has taken over conversations across all sectors of society. In a very short time span, consumers and a wide range of organizations have adopted generative AI technologies with astonishing capabilities (Bove 2023). These new technological developments have accelerated the integration of AI into many tools, apps, and areas of our daily life, but the transformative AI technology is perhaps most deeply impacting the world of advertising. AI-enabled advertising spending worldwide in 2022 was estimated to be $370 billion, with predictions of $1.3 trillion in the next ten years (Statista 2023).

While the viral sensation and enormous popularity of ChatGPT are generating unprecedented attention and interest in AI right now, AI, as both a theoretical concept and version of its technological implementations and applications, has a long history with its origin going back to the 1950s. In the advertising field, the first research article on the topic of AI in connection to advertising was published in 1988 (Cook and Schleede 1988). The authors described “decision support systems” (i.e., expert systems) as probably the “most widely implemented and well-known applications of AI” (48). As described, these systems used databases and models to solve problems, such as help with direct-mail systems and newspaper advertising placement. Since then, AI has changed the business of media and advertising and attracted attention from both advertising practitioners and scholars. Although there is no standard definition, Rodgers (2021) defined AI advertising as “brand communication that uses a range of machine functions that learn to carry out tasks with intent to persuade with input by humans, machines, or both” (2) and positioned AI advertising as a subdiscipline of advertising “situated at the intersection of cognitive science, computer science, and advertising” (2). Advertisers are using AI technologies in automated market segmentation and targeting, ad creative development and personalization, improving ad buying and placement, and optimizing advertising investment (Kietzmann, Paschen, and Treen 2018).

Following the trend of increasing AI adoption in advertising, advertising scholars have organized sessions at the American Academy of Advertising (AAA) conference, such as the 2014 preconference, “Big Data for Advertising Research and Education,” and the joint AAA-ANA Educational Foundation luncheon panel with the provocative title, “The Future of Advertising—Will We Be Replaced by Robots?” in 2018.

The Journal of Advertising (JA) has published multiple themed collections on AI-related topics, starting with a Special Section on Artificial Intelligence and Advertising, guest-edited by Hairong Li (2019). This collection of articles explored the potential and actual application of AI technologies to enhance advertising efficiency, effects, and effectiveness across the entire spectrum of the advertising campaign process, from situation analysis and consumer insight generation to advertising message creation, to media planning and buying, and to advertising effect assessment (see the Journal of Advertising, vol. 48, no. 4).

Another Special Section on Advances in Computational Advertising in 2020, guest edited by Jisu Huh and Ed Malthouse (2020), addressed broad implications of evolving computer science technologies for data-driven, AI-powered computational advertising, and proposed a future research agenda in the areas of macro and exogenous factors, consumers’ roles in computational advertising, AI-powered ad content generation, computational advertising media planning strategy shifts from purchasing exposure to focusing on meaningful consumer engagement, and computational advertising measurement systems (see the Journal of Advertising, vol. 49, no. 4).

A year later, JA published an up-to-date comprehensive Themed Issue on Promises and Perils of Artificial Intelligence and Advertising (2021, vol. 50, no. 1). In her editorial, the previous editor, Shelly Rodgers (2021) proposed an AI classification schema to systematically understand and develop subdomains in the emerging AI advertising research field along the following three dimensions: AI type, AI function, and learning type. The articles in that themed issue examined promising potentials as well as potential risks of AI-powered virtual influencers, context-aware ad/brand placement algorithms, AI-enabled creative advertising systems, computer vision to detect the image–text match and mismatch, AI-enabled in-store communication in an offline shopping environment, and social media listening-platform AI. Again in 2022, included in the JA’s 50th Anniversary Special Section Reimagining Advertising Research: 50 Years and Beyond, Coffin (2022) proposed a maieutic research agenda for future research on AI advertising covering three broad categories of questions dealing with applications of AI technologies to advertising–ontological, technical, and ethical questions of AI advertising.

As these recent special collections and other individual articles showcase, leading scholars and industry thinkers in our field and neighboring disciplines are actively examining and engaging in debates on AI technologies and their applications to advertising practices and effects. However, we have not imagined such powerful AI technologies as ChatGPT emerging and spreading in the general public so quickly. According to industry estimates, ChatGPT reached 100 million monthly users in the first two months after launch, which makes it the fastest-growing technology application in history (Chow 2023). ChatGPT and other generative AI technologies in this new phase of AI advancement are expected to completely transform the advertising business and research. More research is urgently needed to gain an understanding of the short- and long-term impacts of this new generation of transformative AI technologies on advertising across the micro, meso, and macro levels. In this editorial, we attempt to shed light on some of the most important emerging research problems and propose directions for research development on the topic.

AI-Driven Transformations and Emerging Research Problems

We identify four general areas of advertising practice and research that AI is fundamentally altering and that open opportunities and need for research: (1) consumers’ experience of advertising; (2) societal and policy implications related to the truthfulness of AI-generated content; (3) the analytical considerations of data and algorithms; and (4) the functioning of the advertising industry. Figure 1 provides an illustration of these four research areas.

Figure 1. Illustration of AI advertising research areas.

Consumers’ Experience of Advertising in an AI-Enabled World

AI is enabling new forms of interactions between humans and AI-enabled “others,” be it AI-driven chatbots or virtual influencers and spokespersons. These developments are calling for new models to explain and predict human-AI interactions. How will consumers interact with human-like AI persuasion agents—whether they take the form of virtual chatbots, virtual social-media influencers, or virtual brand spokespersons? Whether and how will consumers process and react to information and recommendations presented by AI-powered virtual persuasion agents and human agents differently? What are important influencing factors shaping consumers’ different interactions and reactions (e.g., fear of AI)?

From a sociocultural perspective, AI technologies are challenging our perceptions of reality and stretching the meaning of philosophical notions like hyperreality, whereby reality and simulation of it blur. Will consumers be able to detect and distinguish AI communicators from real human communicators, and how will AI-powered relationships differ from real-world relationships? What might be the societal, cultural, and health implications of interacting in an AI-powered world, especially amongst those generations who will not have known anything but that world?

The rapidly evolving new generation of AI technologies, such as generative AI-powered search apps, is also likely to fundamentally transform search advertising. Currently, search advertising has the largest share of the digital advertising format, taking up more than 40% of the total digital advertising expenditures (Interactive Advertising Bureau 2023). Since Google emerged as the dominant search engine worldwide and the current search advertising model was established more than twenty years ago, how consumers conduct searches as a gateway to the internet and how search advertising is done has gone largely unchanged. However, with the latest generative AI technology, the way consumers search and find information and entertainment is fundamentally changing. Today, while typical Google search engine results pages (SERPs) contain a large number of ads, AI-powered Bing search results contain virtually no advertising (yet). Some key questions begging our research attention include: How might AI technology affect consumers’ search behaviors and the way their search queries are formulated, the way search results are presented and processed and acted on by consumers? What will be the future of search advertising? How will generative AI transform and reshape search advertising?

Finally, AI technology advancement is also heavily influencing social-media influencer marketing, with the rise and increasing prevalence of AI-powered virtual influencers. These advances raise fundamental questions about how generative AI-enabled virtual influencers might change the dynamic of social-influencer marketing and how they impact consumers. Going beyond these two types of advertising, how will the new AI technology change the other types of advertising across all media and the fundamental principles of how consumers are exposed to, experience, process, and react to ads? Will AI-powered new types of advertising compel us to imagine and develop a completely new theory of advertising processes and effects or would the existing advertising theory still hold true?

The Truthfulness of AI-Generated Content and Policy Implications

The applications and impacts of AI in advertising have been discussed and examined primarily from the perspective of the positive potential of AI in improving advertising effects and effectiveness and enhancing the efficiency of the advertising process. However, the recent generative AI technology development raises serious concerns and questions about many potential risks. Going beyond the advertiser-initiated and advertiser-controlled AI advertising and advertiser-expected effects, we should pay attention to the broader social impact and implications of brand-initiated and consumer-initiated AI-generated content that contains fake and even false information. AI-generated text, images, and videos are already spreading online; some are benign (e.g., AI-generated fake pizza restaurant commercials posted on Twitter), while others can be quite harmful to society (e.g., false news stories, harmful deepfake videos). Fake images and videos of virtual or real humans communicating information or entertainment content can contain false information, as demonstrated by many stories of ChatGPT hallucination.

Advertising researchers can tackle these important issues from the ground up, with advertising literacy training, and from the top down, by considering the regulatory implications. Existing models of media and advertising literacy (Nelson 2016) could be adapted to include interventions to help consumers detect and prevent the spread of false information in an AI-enabled world, and researchers can continue to assist in expanding advertising literacy efforts to educate audiences about AI-powered communications. With regard to regulation, technological and legal remedies for dealing with these problems are emerging across computer science and policy/law (e.g., AI-generated text detection tools and the EU’s proposed AI regulations). Multiple research problems specific to advertising emerge from there, including whether and how the advertising industry should address the ethical and legal issues surrounding the possibility of AI-generated advertising content containing false or misleading information as well as strategies for designing and implementing AI-powered advertising in a socially responsible and ethical manner.

Use of Data and Algorithms

In addition to advertising processing and effects and advertising business-related issues, many ethical issues arise with regard to data collection, data harvesting, and algorithm transparency in connection to AI advertising. Given that building AI models requires massive amounts of data, the business practice of collecting, aggregating, and mining increasingly large amounts of consumer data for building AI advertising models calls for more research into the evolving nature of data privacy issues. In addition to rapid expansion and increasing sophistication of surveillance technology, there are shifts in data collection toward more covert data capture without consumers’ awareness or actions, and integration of surveillance data across multiple platforms, devices, and even across offline and online consumer touch points (Strycharz and Segijn 2022). More research is warranted examining questions like: What are the implications of increasing covert data capture vs. overt data for the advertising industry and consumers? How can we explore consumers’ awareness, knowledge, perceptions, and acceptance of expanding types of data collection points and what are implications for advertising practices and effects?

The more advanced AI models become, the more opaque the tools tend to become, and most of today’s AI models operate in a black box lacking explicability and scrutability (Rai 2020). As discussed in previous literature (e.g., Helberger et al. 2020; Hermann 2022), algorithmic transparency is critical to developing ethical and socially responsible AI technology and its use in advertising. The lack of transparency and consumers’ limited ability to understand AI technology and the mechanism and processes of AI-powered advertising would affect advertising outcomes.

Algorithmic bias is another connected issue. Biased and unfair predictions made by AI algorithms used in the digital-media and advertising industries often occur when the training data used to build the computational models do not fully represent the population of consumers for which the models are being built (Noble 2018), which creates algorithmic bias and unfairness. Related research questions could include algorithmic bias in data-driven computational advertising and its impact on society’s inequality problems, and bias or inequities in the current advertising measurement metric system and their impact on the AI-powered advertising message creation and personalization, targeting, and ad buying and placement.

Impact of AI on the Advertising Industry

Expanding the research scope to the economic impact of AI, setting aside the question of the existential threat of AI on the human race, one of the most hotly debated macro-level questions since the release of ChatGPT, or even beforehand, has been whether AI will make human jobs, and which of them, obsolete; to what extent human workforces would be replaced by AI technologies, and what will be the negative economic impacts on individuals and the whole society? Alan Turing (1950), in his seminal article “Computing Machinery and Intelligence,” envisioned a future where “we may hope that machines will eventually compete with man in all purely intellectual fields” (460), whereas another noteworthy mathematician/computer scientist, Lady Ada Lovelace, believed that the analytical engine “has no pretensions to originate anything” (qtd. in Hartree 1949, 70). Although the past advancement of automation and earlier iterations of AI technologies have replaced a lot of human jobs, the advertising industry and its workers, especially creative-sector professionals, appear to be insulated so far from the AI impact, even though AI is changing the way creatives approach their work (Kulp 2023). New advertising jobs have been created because of AI (e.g., the entire tech industry relying on advertising-generated revenues, advertising technology firms, and data analytics jobs).

However, the current wave of generative AI will likely cause a much bigger storm in the advertising businesses and workers. The advertising industry has already started embracing AI to enhance efficiency, generate ideas, develop campaign drafts, and improve ad targeting and personalization (Li 2019). As generative AI technology facilitates the development of content and messaging through prompt engineering—“the process by which a user can fine-tune their inputs to achieve a desired output”—creators across the entertainment and advertising industries are expressing genuine concern over the future of creativity (Short and Short 2023, 1).

Some of the interesting questions include: To what extent will the AI applications continue or expand, and what areas of advertising business and practices will be more vulnerable to AI replacement, and which areas will still require human labor? Under what conditions would AI replace creative professionals? Which advertising tasks will still require human creativity, supervision, and intervention? How will AI change advertising companies’ revenue models and organizational structure and the shape of their workforces? How will AI affect and change the relationship between clients and advertising agencies? What is the meaning of creativity in an AI-powered environment?

What the evolution of AI technologies will bring to advertising is anyone’s imagination. What we propose here is in no way a thorough list of potentially important research questions, and they are meant to serve as thought starters. Our intention is to draw your attention to an important avenue of transformative research to contribute to the rapidly evolving discussion of the psychological, sociological, and cultural implications of an AI-powered advertising world. We also call advertising researchers’ attention to the needs for novel educational and regulatory frameworks to guide the future of AI technologies and the future of AI advertising so that advertising can continue to serve all stakeholders and function as an important social institution that brings innovative solutions to problems. We hope that this editorial, along with the prior research articles on AI and advertising, stimulates and inspires forward-looking research and systematic theory development.

Generative AI’s Implications for Academic Research and Publication Policies

As we encourage future research in this area, we cannot ignore the disruptive impacts and emerging problems we face as academic researchers stemming from the new generative AI tools. Thus, this section addresses the issues related to the impacts and applications of generative AI in how we, academic researchers, conduct, produce, and disseminate our work.

ChatGPT and other recent generative AI tools are prompting academic associations (e.g., International Conference on Machine Learning [ICML], Association for Computational Linguistics [ACL]) and journals to update their policies or create new ChatGPT-related policies to deal with potential research-ethics issues, such as potential plagiarism, false information, and authorship issues. For example, the ACL 2023 Policy on AI Writing Assistance (https://2023.aclweb.org/blog/ACL-2023-policy/) provides specific guidance about the use of generative AI models, including “If such tools were used in any way, the authors must elaborate on the scope and nature of their use.”

Several journals have also announced policies regarding the use of ChatGPT or any other generative AI tools in scholarly writing. Among the strictest is an updated policy by Science journals (https://www.science.org/content/page/science-journals-editorial-policies). Its editorial policies are now updated with the following statement:

Text generated from AI, machine learning, or similar algorithmic tools cannot be used in papers published in Science journals, nor can the accompanying figures, images, or graphics be the products of such tools, without explicit permission from the editors. In addition, an AI program cannot be an author of a Science journal paper. A violation of this policy constitutes scientific misconduct.

Other journals’ policies tend to be less strict and allow the use of AI tools as long as such use is disclosed. For example, Proceedings of National Academy of Sciences of the United States of America (PNAS)’s policy (https://www.pnas.org/post/update/pnas-policy-for-chatgpt-generative-ai) states:

If AI software such as ChatGPT has been used to help generate any part of the work it must be clearly acknowledged; it must be noted in the Materials and Methods section (or Acknowledgments, if no Materials and Methods section is available) on submission. The software cannot be listed as an author because it does not meet the criteria for authorship and cannot share responsibility for the paper or be held accountable for the integrity of the data reported.

Based on careful consideration of this trend and the guidelines of our current publisher, Taylor & Francis, the Journal of Advertising has established a new policy regarding the ethical use of AI as follows:

Large language models (LLMs) and AI tools do not meet the criteria for authorship and so cannot be listed as an author. Authors are responsible for the originality, integrity, and validity of the content of their submissions and need to be able to enter into an author publishing agreement. Use of such tools in the writing of an article must be done responsibly and transparently in accordance with publishing ethics guidelines.

This new guidance is included in the new Journal of Advertising Research Quality and Ethics Guidelines that are presented in the next article. This guidance mainly addresses the authorship-related ethical issue. The norms, ethical standards, and policies about the use of ChatGPT and other generative AI tools in conducting research (including literature review) and writing manuscripts are still developing across disciplines. However, as a general principle, we would discourage the use of AI tools for conducting a literature review, summarizing and synthesizing literature, and scholarly writing in submissions to the JA, with the exception of using AI editing tools for proofreading, grammar check and correction, and copyediting a draft written by human researchers.

As technology and advertising keep changing and pose new challenges and opportunities, the field of research on technology-driven or technology-powered advertising tends to be dominated by practitioner research driven by technical questions. As noted in Coffin (2022), “at this relatively early stage in the evolution of AI advertising, it seems a suitable time to call for a more reflective, maieutic approach” (617) to examining AI and advertising. We call for such research promoting the use of AI for ethical and responsible advertising and social good, and we look forward to more creative and innovative research informing advertising theory development and guiding advertising practice.

Jisu Huh
Editor-in-Chief
Michelle R. Nelson
Senior Associate Editor
Cristel Antonia Russell
Senior Associate Editor