Editorial: The changing landscape of marketing research in the AI era: prospects and challenges

What the paper says

The advent of artificial intelligence (AI) – encompassing both generative and analytical AI – has dramatically transformed the landscape of marketing research and practice, spanning theoretical frameworks, cutting-edge research frontiers, methodological tools and ethical guidelines. Driven by AI and recommendation algorithms, interactive marketing domains and research realms have transcended the scope of digital marketing tools, customized data collection, customer connection, engagement and participation (Wang, 2021; Yu, 2023), shifting toward sophisticated market predictive analytics, precise forecasting, in-depth consumer insights and real-time hyper-personalized recommendations (Habil et al., 2023). For instance, natural language processing (NLP) is deployed for sentiment analysis on vast volumes of unstructured text data derived from social media, customer reviews, discussion forums and consumer surveys. This enables researchers to gauge public sentiment, identify customer pain points and detect emerging trends in outcome evaluation – including the measurement of consumer brand perception and interactive experiences with virtual influencers (Nghiêm-Phú and Suter, 2023). In this way, AI-powered marketing research unlocks rich qualitative insights from quantitative datasets, establishing real-time feedback loops that inform product development and marketing strategy formulation throughout interactive processes.Meanwhile, the rapid evolution of generative AI, coupled with the proliferation of AI replicas (e.g. digital doppelgängers, digital twins and digital personas) and deepfake technologies, entails inherent risks. These include database contamination (Burden et al., 2025), AI hallucinations (Wen and Laporte, 2025), model collapse and the average trap (Huang and Rust, 2025), as well as ethical dilemmas such as algorithmic bias, identity theft and privacy and security concerns (Grewal et al., 2025). Collectively, these issues underscore the double-edged nature of AI, posing profound challenges for both academic inquiry and practical implementation.This article first provides an overview of the evolving research paradigms and emerging themes in marketing research during the AI era. It then outlines theoretical reconceptualizations and methodological advancements catalyzed by AI technologies. Finally, it examines the challenges and potential risks associated with AI tool applications in marketing research and practice and delineates future research directions.Advancements in human-machine interaction have blurred the boundaries between human input and machine output, driving a paradigm shift in marketing research. Traditional interactive marketing frameworks – centered on human-human, human-brand and human-technology interactions (Wang, 2021, 2023) – have given way to an “algorithmic symbiosis” model, which emphasizes synergistic collaboration between humans and AI-based systems (Almeida and Senapati, 2024; Wang). Algorithmic symbiosis denotes a mutually beneficial coexistence of human and machine intelligence, wherein both entities collaborate to enhance each other's capabilities (Litvinova et al., 2024).This paradigm shift necessitates a re-examination of the theoretical debate surrounding AI'srole: whether it functions as an assistant, substitute, replacement or complement to human agents. Early studies explored how voice-activated assistants and AI-driven chatbots optimize customer inquiry responses by reducing latency and improving operational efficiency (Candao et al., 2023). As AI technologies have grown more sophisticated, scholarly attention has shifted from AI'sauxiliary role to its function as an empowered agent that actively participates in value co-creation processes (Wang, 2024). In contemporary marketing practice, AI systems are no longer viewed merely as tools or assistants but as extensions of human agents within synergistic partnerships – leveraging the complementary strengths of human intuition and algorithmic precision. Crucially, human-AI collaboration requires the integration of human judgment and insight with machine-driven accuracy, ensuring mutual augmentation rather than the replacement or diminishment of human agency and creativity (Gonzalez, 2025). Thus, the symbiotic framework of human-machine intelligence lays the groundwork for future interactive marketing research, enabling scholars to derive actionable insights into market dynamics.For decades, marketing and consumer psychology research has been rooted in traditional cognitive and affective response paradigms, drawing on information processing theory, the stimulus-organism-behavior (S-O-B) model and planned behavior frameworks. These approaches typically use consumer attitudes and purchase intentions as proxies for actual buying behavior, seeking to explain consumer actions through individual beliefs, values, emotions and social norms. However, empirical evidence indicates that the correlation between self-reported intentions and real-world behavior is relatively weak to moderate (Rhodes and Dickau, 2012).The advent of AI-driven automation, big data collection and data mining technologies – capable of aggregating information from corporate databases, Internet search histories, purchasing records and social media behaviors – has equipped consumer researchers with tools to measure behavioral outcomes directly as dependent variables. This trend reflects a resurgence of neobehaviorist thought, which emphasizes the empirical observation of overt behavior while also incorporating mental events as theoretically explanatory mechanisms from a logical positivist perspective (Greenwood, 2015). Consequently, there is a growing demand for integrated research designs that combine primary data from controlled experiments (to ensure theoretical rigor and internal validity) with real-world business data (to capture behavioral outcomes and ecological validity) (Wang, 2025).As AI has emerged as a central actor in marketing and consumption systems, it has fundamentally reshaped the consumer journey and transformed theoretical understandings of how individuals search for, evaluate and select products. Machine learning algorithms capture and analyze consumer behaviors in real time, creating highly automated and adaptive marketing ecosystems (Mariani et al., 2022). This has necessitated a reconsideration of the theoretical foundations of consumer behavior analysis, particularly regarding how AI-driven personalization, algorithmic decision-making and AI-generated content reshape existing research frameworks and bridge the gap between theory and practice.Notably, analytical AI has redirected theoretical focus from descriptive analyses and phenomenon explanation to the anticipation of consumer behavior via sophisticated predictive modeling. This enables dynamic pricing adjustments and real-time optimization of market performance based on consumer sentiment and demand fluctuations. Instead of reacting to past consumer actions, AI algorithms proactively forecast future purchase intentions, delivering highly individualized content, product recommendations, offers and customer service interactions across multiple touchpoints.The power of hyper-personalization – facilitated by real-time predictions based on consumer response behaviors and contextual data – has expanded and enriched the theoretical underpinnings of market segmentation, targeting and positioning (STP) strategies. AI enables one-to-one marketing by analyzing large datasets to identify precise customer segments, transforming traditional STP practices through predictive analytics and cross-channel real-time personalization. As a result, AI technologies have advanced theoretical understanding and practical applications of core marketing constructs in the following key ways:Ultimately, AI generates customized, interactive marketing materials tailored to individual preferences – including product features, messaging, targeted promotions and behavioral pricing – to meet the unique needs of consumers, replacing one-size-fits-all value propositions and product solutions via interactive marketing communication.Rapidly evolving AI technologies continuously generate new research questions and drive the exploration of emerging phenomena, resulting in a shortened “research product lifecycle” – defined as the period from a research theme's initial development to its maturation and decline. Whereas traditional marketing constructs or research topics often remained relevant for decades, many contemporary AI-related topics exhibit much shorter lifecycles: “trendy” themes can quickly become obsolete as technological advancements redefine academic frontiers.For example, NLP has facilitated the widespread adoption of chatbots and sentiment analysis, spurring a proliferation of empirical studies across diverse marketing contexts that are approaching maturity at an accelerated pace. This demands that the academic community continuously redefine cutting-edge research agendas and explore new frontiers with agility and originality, keeping pace with exponential technological progress and theoretical breakthroughs. Moreover, “older” research topics require revisiting and re-evaluation through novel theoretical lenses and alternative perspectives, generating fresh insights to inform marketing strategy and decision-making processes.Interactive AI enables efficient communication automation by integrating human behavior and intelligence into machines and systems. The adoption of both generative and analytical AI has driven the evolution of research methodologies, leading to rapid advancements in AI-augmented research methods across multiple dimensions.AI has unlocked exponential growth in data processing capabilities, enabling the analysis of large-scale databases that were previously unmanageable. This has expanded research capabilities beyond traditional time-consuming and labor-intensive methods, allowing researchers to analyze customer purchase histories, browsing behaviors, social media interactions and customer relationship management (CRM) records to create highly personalized customer profiles. AI-powered research is uniquely capable of uncovering hidden relationships, patterns, trends and potential behavioral responses that were previously inaccessible, shifting the research paradigm from hypothesis testing on small samples or curated datasets to big data analysis and synthesized data-driven predictive modeling.Even consumer research – traditionally reliant on primary data collection via scenario-based laboratory experiments – now increasingly incorporates large secondary datasets to capture real-world consumer behavioral outcomes, enhancing the ecological validity of findings.Traditional qualitative research methods are often susceptible to interpretive bias and limited generalizability, restricting their mainstream application in marketing research despite their ability to uncover deep, contextual understandings of phenomena and provide rich insights into consumer experiences, motivations and behaviors that may be overlooked by aggregated quantitative data.AI-enabled tools – such as machine learning algorithms and NLP-based semantic analysis – have transformed the way marketing researchers conceptualize hypotheses, conduct studies and interpret results. By combining quantitative analysis (to identify patterns in large datasets) with qualitative insights (from focus groups, in-depth interviews, ethnographies and unstructured data sources like customer reviews and social media posts), AI applications facilitate a more nuanced understanding of marketing phenomena. This integrated approach enables researchers to uncover consumer emotional drivers, cognitive patterns and contextual influences faster and more efficiently than traditional methods.AI algorithms enable the synthesis of multi-modal data, facilitating the simultaneous analysis of diverse and complex multimedia formats (text, images, audio and video) and rich sentiment data. These algorithms also support rigorous validation of sentiment scores, capturing real-time consumer engagement and refining the measurement of attitudinal constructs (e.g. brand loyalty, emotional intensity, perceived value and review helpfulness) through automated extraction and scoring of emotions and nuanced linguistic tones.Furthermore, AI algorithms integrate online data (e.g. browsing history and product reviews) with offline data (e.g. in-store visits and shopping experiences) to generate comprehensive assessments of consumer attitudes and behaviors. This holistic approach provides cross-platform insights across social media, physical stores, e-commerce and social commerce channels, enabling the orchestration of seamless omnichannel customer journeys across all touchpoints.Content analysis has long been a staple of marketing research – particularly in advertising studies – but its use has declined due to its time and resource-intensive nature. Traditional content analysis is also plagued by inherent limitations, including small sample sizes, generalizability issues, potential subjectivity and bias in manual coding and over-reliance on manifest content that overlooks nuanced aspects of communication (e.g. tone and implied meaning).AI and transformer models, or computational content analysis (Barari and Eisend, 2024), address these limitations by enabling automated extraction of meaning, emotion and intent from large volumes of unstructured text via sentiment analysis, topic modeling, entity recognition and intent recognition. AI-powered content analysis transcends the constraints of manual coding, processing massive datasets at scale while delivering more granular, accurate and nuanced insights into consumer feelings and perceptions by capturing latent content and contextual nuances.Generative AI-based simulation and modeling techniques – capable of embedding human behavior and intelligence into machines and systems – allow researchers to create virtual representations of complex marketing systems. These tools enable the simulation of diverse market scenarios and consumer profiles to predict potential outcomes of consumer responses and marketing strategies (Sarker, 2022).For example:In this way, generative AI transforms market research by leveraging synthetic data to simulate market dynamics, design experimental scenarios and test theoretical frameworks through agent-based modeling with AI-driven decision-making rules (Korst et al., 2025). Ultimately, generative AI has the potential to reduce reliance on human samples by simulating virtual consumers that mimic real-world customer behavior. Recent research demonstrates that generative AI can effectively replicate classic behavioral games designed to elicit human traits (e.g. trust, fairness, risk aversion and cooperation) and generate responses to psychological personality surveys that are indistinguishable from those of large random human samples (Mei et al., 2024).Alongside the rapid advancement of AI technologies come inherent limitations that present necessary trade-offs. The transformative impact of AI on marketing research and practice demands ethical and responsible scrutiny, particularly regarding how algorithmic bias influences marketing decision-making, predictive recommendations and consumer choice behavior.The growing use of large language models (LLMs) and other generative AI tools is contaminating the online data ecosystem, primarily through model collapse and synthetic data pollution. When AI systems are trained on synthetic content, they produce increasingly degraded outputs as erroneous data is inadvertently incorporated into training datasets. Additionally, data poisoning attacks may occur when malicious actors intentionally introduce harmful or deceptive data into AI training datasets to skew model outputs in specific directions (Burden et al., 2025). As AI training models become increasingly self-referential – entering a “self-consuming loop” – their performance deteriorates, and they gradually lose alignment with human preferences while amplifying existing biases (Huang and Rust, 2025).To mitigate risks of identity theft, privacy breaches and discrimination, many organizations use de-identification or pseudonymization techniques when processing big data. However, these practices often lead AI algorithms to make predictions and decisions at the group level, focusing on population-level patterns, trends and probabilities rather than individual characteristics (Mittelstadt, 2017). As a result, individuals are primarily treated as average members of population groups based on perceived “typical” behavioral patterns, causing unique individual traits to be lost or aggregated within predictive models. This phenomenon, known as the “average trap,” results in generative AI-generated content that only predicts the most likely but uninspired scenarios (Huang and Rust, 2025).Consequently, generative AI-simulated market scenarios and virtual consumers may compromise the predictive accuracy and nuanced assessment of diverse individuals and market segments in marketing research. In consumer behavior studies, generative AI-generated responses often lack granularity and depth, highlighting the need for future research to examine how AI-designed research stimuli, experimental protocols and simulated market responses – based on synthetic data generation – truly reflect the complexity and diversity of real-world consumer behavior, along with associated technical and ethical caveats.Hyper-personalized algorithmic recommendations and targeting may reinforce the formation of “information cocoons” (Sunstein, 2006), wherein individuals are only exposed to content aligned with their existing preferences and beliefs. This creates information gaps between consumer groups with different interests, narrowing the range of information accessible to individuals and exacerbating cognitive differences across groups (Matz, 2021).Recent psychiatric research has linked AI chatbot usage to potential psychosis risks, as iterative conversations between humans and AI companions may lead to shared delusions (Schechner and Jargon, 2025). Chatbots like ChatGPT often prioritize user validation by echoing their beliefs, accepting user thoughts as factual truth and creating a feedback loop that amplifies delusional thinking, potentially contributing to mental health issues.The rise of AI-driven big data analysis has shifted the focus of marketing research from explaining why consumers behave a certain way to describing what they do and how they behave. This approach emphasizes observed market performance and direct behavioral responses – for example, measuring influencer conversion rates in livestream commerce, analyzing the correlation between social media likes and clicks or patterns marketing strategies to purchase rates or reliance on data-driven analysis the traditional from which to and from psychological inquiry (Greenwood, 2015). This in AI-driven marketing research as over-reliance on AI may human theoretical understanding and contextual approach that AI analytics with human judgment is to the of AI automation in marketing research. through theoretically designed studies and primary data collection (e.g. consumer surveys and can test mechanisms and uncovering and beyond the patterns in data (Wang, are trained by vast datasets to identify patterns and generate new content by the most of This training AI algorithms to – the generation of or content et al., – often from training data or inherent model limitations and 2023). content may include events or academic et al., 2023). For example, a that generates article in of to for 2025). AI tools may existing data on a topic and create and the of generative AI, deepfake and other digital replicas enables the of highly to identity and theft that both and widespread adoption of AI for content has concerns within the academic community regarding and have and ethical the use of AI tools in and key has the use of generative AI tools for content and reviews) is while to language and is example, that AI tools be in an responsible and human throughout the rules to be treated as and are from or to generative AI tools, which may and data privacy on human interpretive for the accuracy, rigor and validity of their review of AI offers but also ethical challenges that the of marketing research and AI-driven marketing research on the collection of vast volumes of data, including social media Internet data, purchasing histories, online and recognition data. This data collection risks of privacy data breaches and identity consumers lose in the and of algorithms, they are likely to or information or may provide or data during AI-based data AI models trained on data that reflect population patterns may or existing biases 2024), and generating algorithmic recommendations that or customer the development and of AI – which provides of how AI models generate marketing recommendations – has become and algorithmic fairness, and consumer and participation in marketing research et al., AI has transformed marketing research through theoretical – shifting from human-AI interaction to symbiotic collaboration – and by integrating AI automation and analytics with human intuition and advancements have the way for combining AI-powered big data analysis with insights from qualitative data collection, data-driven outcome analysis with theoretically explanatory expanded research landscape has new to and traditional marketing constructs through novel evolving and research and the and challenges of AI in marketing research is to academic The future and responsible use of generative and analytical AI on a understanding of their limitations and the of associated risks – including data algorithmic bias, data privacy concerns and