From Ad Algorithms to Consumer Insights: The Evolution of AI Advertising

By Samudrapom DamReviewed by Susha Cheriyedath, M.Sc.Jul 5 2023

In an article in press with the Journal of Business Research, researchers reviewed artificial intelligence (AI) advertising literature to understand the evolution of AI advertising research and present an overview of this field for future research.

Background

Advertising has rapidly evolved in recent years due to the rising adoption of AI and its applications. AI has made advertising more intelligent, targeted, personalized, and competent by facilitating and automating crucial advertising functions such as impact evaluation, advertisement creation, buying, media planning, and consumer insight discovery.

Advanced advertising media use AI to optimize ad delivery and improve advertisement effectiveness. AI employs image and speech generation, natural language generation (NLG), machine learning (ML), speech recognition (SR), image recognition (IR), and natural language processing (NLP) to assist advertisers in different advertising functions, including automated ad generation and ad optimization.

Programmatic and computational advertising utilize AI to develop targeted promotions by enabling automated ad placement, scheduling, and media planning and buying. Additionally, the application and role of AI-enabled smart bots and speakers are increasing in marketing. For instance, marketers use AI-powered smart speakers, such as Google Assistant, Siri, and Alexa, to interact and engage with target audiences and innovatively deliver promotional messages.

Similarly, brands employ AI chatbots with ML and natural language understanding (NLU) to determine customer requirements, improve AI-consumer interaction, understand customer responses, and provide customer support.

AI advertising also plays a critical role in generating a higher return on investment (RoI) for advertisers by improving the personalization and targeting of advertising, which makes the ad more efficient and effective. Moreover, AI and generative adversarial networks (GANs) are used to transform the video or image of people in advertisements.

Review of AI advertising literature

In recent years, research on AI advertising has provided several insights into its potential features, functions, uses, and challenges. However, studies have yet to offer a comprehensive AI advertising literature overview that can provide research directions for practitioners and scholars.

In this paper, the authors reviewed AI advertising literature to provide a detailed understanding of this field and identify future research areas. The evolution of AI advertising research was mapped by performing a framework-based and bibliometric analysis of 75 AI advertising research articles published from 1990 to 2022.

Researchers performed a bibliometric analysis for science mapping, bibliographic coupling, and performance analysis of the AI advertising literature to identify the research focus and key trends.

The theory context characteristics method (TCCM) framework was applied to provide a structured and comprehensive representation of AI advertising research and its contribution to AI advertising theory development.

Significance of the study

The key findings of this study were the AI advertising publication trends, research contexts identified by bibliographic coupling, and TCCM classification. Systematic literature analysis reduced the possibility of subjective bias and provided a comprehensive overview of AI advertising research.

The AI advertising research evolution was analyzed in two phases: the first from 1990 to 2017 and the second from 2018 to 2022. In the first phase, 14 journal articles on AI advertising were published that primarily focused on using AI algorithms and expert systems for scheduling, media planning, targeted promotions, and personalization.

During the second phase, 59 articles were published within five years, indicating increased interest in using AI in advertising. These articles focused on advertisement campaign optimization, programmatic advertising, advertisement creativity, ad performance measurement and evaluation, and ad delivery.

Theoretically, grounded research was rare in AI advertising, with only 10 articles published in recent years that have employed theories. Six clusters of ideas were employed extensively in AI advertising literature, including technology, sociology, psychology, media, mass communication, and computer science. Crucial insights into user concerns, consumer adoption, and ad effectiveness were gained using these theories.

Advertisement optimization was the most recognized context in the AI advertising literature, with 18 studies investigating AI’s role in improving ad effectiveness. However, other essential contexts such as ad creativity and ad generation received less attention in the literature.

Conceptual articles were the most dominant in the AI advertising literature, with 72 articles being conceptual out of 74 studies. These articles played a crucial role in driving new theoretical development. The computational research method was the popular research design in AI advertising literature. This empirical approach employs experiments on datasets through expert systems or algorithms.

Four themes emerged as the key AI advertising research focus areas: advertising effectiveness, computational advertising, AI-driven innovations in advertising, and programmatic automation and advertising.

Conclusion and future outlook

To summarize, the findings of this review in AI advertising will help researchers/authors to effectively map the evolution of AI advertising literature and assess the intellectual, thematic, and conceptual structures of this field. In the future, AI advertising research must consider strategies to integrate and incorporate AI with existing marketing functions and emerging functions, such as blockchain.

Additionally, more research is required to determine consumer attitudes and perceptions toward AI-delivered ad functions, achieve better personalization without affecting consumer privacy through secure and transparent processes using computational methods, and identify ways to improve programmatic advertising functions and advertisement effectiveness.