The Impact of Generative Artificial Intelligence in Advertising on Multi-Generational Purchase Intentions: An Empirical Study in China ()
1. Introduction
The rapid advancement of digital technologies has fundamentally reshaped how firms communicate and interact with consumers. Among recent technological developments, artificial intelligence (AI) has emerged as one of the most transformative forces in digital marketing. By enabling large-scale consumer data analysis, automated content generation, and personalized communication, AI has become deeply embedded in contemporary marketing practices (Huang & Rust, 2021; Kaplan & Haenlein, 2019; Dwivedi et al., 2021; Dwivedi et al., 2023). Applications such as AI-generated advertising visuals, automated copywriting, intelligent recommendation systems, and virtual brand representatives are now widely implemented across digital platforms (Campbell & Farrell, 2020).
As AI technologies expand within the marketing domain, consumer decision-making processes have become increasingly complex. Consumers are continuously exposed to algorithm-driven content across social media, e-commerce platforms, and short-video applications (Xu & Pratt, 2018). AI-generated advertising leverages behavioral data, browsing histories, and preference signals to deliver highly targeted and interactive messages. This shift has altered how consumers evaluate advertising credibility, emotional appeal, and ultimately their purchase intentions (Kim & Kim, 2020; Longoni et al., 2022).
However, responses to technological innovation are not uniform across age groups. Generational cohort theory suggests that individuals shaped by different socio-technological environments develop distinct attitudes toward innovation and digital systems. Generation Z, often described as digital natives, grew up in highly interactive digital environments and tends to exhibit stronger acceptance of algorithmic personalization and virtual communication. Millennials experienced the transition from traditional media to digital platforms and often display a balanced perspective that integrates openness to innovation with rational evaluation (Priporas et al., 2017; Zhang et al., 2018). In contrast, Generation X formed consumption habits prior to the widespread adoption of intelligent technologies and typically emphasizes reliability, transparency, and informational authenticity.
AI-generated advertising influences consumers through both cognitive and affective mechanisms. From a cognitive perspective, perceived credibility plays a central role in shaping trust and reducing uncertainty. Consumers evaluate whether AI-generated content is reliable, accurate, and aligned with brand consistency before forming purchase intentions (Ajzen, 1991). From an affective perspective, emotional resonance reflects the extent to which AI-generated advertising evokes psychological connection and engagement. Features such as anthropomorphic avatars, immersive storytelling, and interactive formats may enhance emotional responses, particularly among younger cohorts.
Existing research on AI in marketing has largely focused on technology acceptance and general consumer attitudes. While prior studies acknowledge the importance of trust, personalization, and emotional engagement, limited research has systematically examined the mediating roles of perceived credibility and emotional resonance within a unified conceptual framework (Rafaeli et al., 2024). Moreover, comparative investigations across multiple generational cohorts remain scarce. This gap restricts a comprehensive understanding of how persuasion pathways differ across age groups in the context of AI-generated advertising.
Within China’s rapidly evolving digital economy, AI-generated advertising has become especially prominent. Urban consumers are frequently exposed to AI-driven promotional content through digital ecosystems that integrate social media, e-commerce, and mobile applications. This context provides an appropriate empirical setting for examining generational differences in psychological responses to AI-enabled advertising strategies.
Building on generational cohort theory and advertising persuasion theory, this study develops a conceptual framework in which AI-generated advertising features influence purchase intention through perceived credibility and emotional resonance. Furthermore, generational cohort is proposed as a boundary condition affecting the strength of these relationships.
Accordingly, the study proposes the following hypotheses:
H1: AI-generated advertising features positively influence perceived credibility.
H2: AI-generated advertising features positively influence emotional resonance.
H3: Perceived credibility positively influences purchase intention.
H4: Emotional resonance positively influences purchase intention.
H5: Perceived credibility mediates the relationship between AI-generated advertising features and purchase intention.
H6: Emotional resonance mediates the relationship between AI-generated advertising features and purchase intention.
H7: The relationships among AI-generated advertising features, perceived credibility, emotional resonance, and purchase intention differ significantly across generational cohorts.
In this study, AI-generated advertising refers to advertising content created or substantially assisted by algorithmic systems, including generative visuals, automated copywriting, and virtual brand representations.
2. Method
This study adopted a quantitative research approach using a survey method to examine the effects of AI-generated advertising on consumers’ purchase intention across different generational cohorts. A quantitative design was selected because it allows systematic examination of relationships among variables through numerical data and statistical techniques, enabling objective hypothesis testing (Hair et al., 2019). The analytical framework of the study is grounded in generational cohort theory and the Technology Acceptance Model (TAM), which together provide a theoretical basis for explaining variations in consumer responses to AI-generated advertising content (Davis, 1989; Venkatesh et al., 2003; Kietzmann et al., 2018).
The research was designed as explanatory research, focusing on identifying and testing relationships among AI-generated advertising features, perceived credibility, emotional resonance, and purchase intention. The study empirically tested a conceptual model derived from prior research in AI-driven marketing and consumer behavior (Rodgers & Nguyen, 2022; Wang & Qiu, 2024).
Data were collected through a structured questionnaire, which served as the primary research instrument. Measurement items were adapted from established scales in existing literature and refined to fit the context of AI-generated advertising. All variables were measured using a seven-point Likert scale ranging from 1 (“strongly disagree”) to 7 (“strongly agree”). The questionnaire comprised four primary constructs: AI-generated advertising features (including innovation, anthropomorphism, transparency, and information quality), perceived credibility, emotional resonance, and purchase intention. A pilot test was conducted prior to formal data collection to ensure clarity and content validity.
The study measured four primary constructs. AI-generated advertising features refer to consumers’ perceptions of technological and creative characteristics of AI-driven advertisements, including innovation, anthropomorphic elements, transparency, and information quality. Perceived credibility reflects consumers’ evaluation of the trustworthiness and reliability of AI-generated advertising content. Emotional resonance captures the extent to which AI-generated advertising evokes affective engagement and psychological connection with the brand. Purchase intention represents consumers’ likelihood of purchasing the advertised product after exposure to AI-generated advertising. Measurement items were adapted from established scales in prior literature and modified to fit the AI advertising context.
Participants were recruited through an online survey distribution approach targeting urban consumers in China. Screening questions were included to ensure that respondents had prior exposure to AI-generated advertising content. Participation was voluntary and anonymous. Responses with excessive missing data or patterned answering were excluded during the data-cleaning process to ensure data quality.
The unit of analysis was individual consumers who had prior exposure to AI-generated advertising through digital platforms, including social media, e-commerce websites, and short-video applications. The target population consisted of urban consumers in China aged between 18 and 65 years. To examine generational differences, respondents were categorized into three cohorts based on commonly adopted generational definitions: Generation X (1965-1979), Millennials (1980-1994), and Generation Z (1995-2009). After removing incomplete and invalid responses, a total of 200 valid questionnaires were retained for subsequent analysis (Generation X: n = 62; Millennials: n = 74; Generation Z: n = 64).
A cross-sectional research design was employed, as data were collected at a single point in time to capture respondents’ current perceptions and behavioral intentions toward AI-generated advertising.
Data Analysis
Data analysis was conducted in multiple stages. First, descriptive statistical analysis was used to summarize respondent demographics and key perception variables. Second, reliability assessments were performed to evaluate internal consistency of the measurement scales using Cronbach’s alpha coefficients.
Third, hierarchical multiple regression analyses were conducted to test the proposed relationships. In Step 1, control variables (e.g., gender, education, and digital media usage) were entered. In Step 2, AI-generated advertising features were included as independent variables predicting perceived credibility and emotional resonance. In Step 3, perceived credibility and emotional resonance were entered to examine their effects on purchase intention and to assess potential mediation effects. Changes in explained variance (ΔR2) were used to evaluate incremental explanatory power across regression models.
To examine generational differences, separate regression analyses were conducted for Generation X, Millennials, and Generation Z. Standardized regression coefficients were compared across cohorts to identify variations in the relative importance of perceived credibility and emotional resonance in predicting purchase intention.
To minimize potential common method bias, several procedural remedies were implemented. Respondents were assured of anonymity, and items measuring independent and dependent variables were presented in randomized order. Additionally, Harman’s single-factor test was conducted to assess common method variance. The results indicated that no single factor accounted for the majority of variance, suggesting that common method bias is unlikely to substantially affect the findings.
3. Results
3.1. Sample Profile
To ensure the representativeness of the dataset and clarify the respondent structure used for hypothesis testing, respondent demographic information was summarized (Tables 1-6).
Table 1. Generational distribution of respondents.
Sample |
Number of respondents |
Percentage |
1965-1979 (Generation X) |
62 |
31.00% |
1980-1994 (Millennials) |
74 |
37.00% |
1995-2009 (Generation Z) |
64 |
32.00% |
Table 2. Gender distribution of respondents.
Sample |
Number of respondents |
Percentage |
Male |
76 |
38.50% |
Female |
82 |
41.00% |
Other |
23 |
11.00% |
Prefer not to disclose |
19 |
9.50% |
Table 3. Distribution of respondents “Education Level”.
Sample |
Number of respondents |
Percentage |
High school/Below |
47 |
23.50% |
Junior college/Associate |
58 |
29.00% |
Bachelor’s Degree |
67 |
33.50% |
Master’s degree |
28 |
14.00% |
Table 4. Regional distribution of respondents.
Sample |
Number of respondents |
Percentage |
First-tier cities (e.g., Beijing, Shanghai, Guangzhou, Shenzhen) |
71 |
35.50% |
Second-tier cities |
63 |
31.50% |
Third tier and lower tier cities |
48 |
24.00% |
Rural areas |
18 |
9.00% |
Table 5. Daily digital media usage time.
Sample |
Number of respondents |
Percentage |
Less than 1 hour |
17 |
8.50% |
1 - 3 hours |
56 |
28.00% |
3 - 5 hours |
69 |
34.50% |
More than 5 hours |
58 |
29.00% |
Table 6. Awareness of AI-Generated advertising.
Sample |
Number of respondents |
Percentage |
Yes, I noticed it |
88 |
44.00% |
No, I did not notice it |
66 |
33.00% |
Not sure |
46 |
23.00% |
These descriptive statistics confirm that the sample encompasses a broad range of age cohorts and consumption contexts, thereby providing an appropriate empirical foundation for examining the generational differences proposed in the research framework. In particular, the distribution of daily digital media usage and the level of awareness of AI-generated advertising indicate that respondents are sufficiently exposed to contemporary digital advertising environments. This level of exposure constitutes a prerequisite for meaningfully evaluating perceptions of AI-generated advertising content.
3.2. Descriptive Findings on Perceptions of AI-Generated
Advertising
Subsequent analysis focused on consumers’ evaluations of AI-generated advertising across several key dimensions, including perceived authenticity, creativity and attractiveness, credibility, emotional resonance, and technology acceptance. Examining these perceptions at the descriptive level is essential, as it establishes baseline consumer attitudes toward AI-generated advertising prior to testing the hypothesized relationships using hierarchical regression analysis.
3.2.1. Authenticity/Naturalness of AI Ads
The results in Tables 7-10 related to authenticity focus on consumers’ evaluations of the realism and naturalness of AI-generated advertising content, as well as the extent to which disclosure of AI generation influences these judgments. Taken together, these findings provide an initial descriptive overview of how consumers perceive the authenticity of AI-generated advertisements before considering downstream effects. The distributions indicate varying levels of perceived naturalness and comfort, suggesting that authenticity remains a salient dimension in consumer evaluations of AI-generated advertising.
Table 7. Distribution of ratings for the naturalness of ai advertising visuals or language.
Sample |
Number of respondents |
Percentage |
Very unnatural |
23 |
11.50% |
Somewhat unnatural |
37 |
18.50% |
Neutral |
62 |
31.00% |
Somewhat natural |
48 |
24.00% |
Very natural |
30 |
15.00% |
Table 8. Evaluation of the realism of virtual characters’ behaviors and facial expressions in AI advertising.
Sample |
Number of respondents |
Percentage |
Completely unrealistic |
19 |
9.50% |
Somewhat unrealistic |
44 |
22.00% |
Neutral |
58 |
29.00% |
Somewhat realistic |
51 |
25.50% |
Very realistic |
28 |
14.00% |
Table 9. Frequency distribution of discomfort triggered by AI advertising.
Sample |
Number of respondents |
Percentage |
Very often |
17 |
8.50% |
Sometimes |
39 |
19.50% |
Occasionally |
74 |
37.00% |
Never |
70 |
35.00% |
Table 10. Impact of “AI-Generated” labeling on perceived content authenticity.
Sample |
Number of respondents |
Percentage |
Significantly increased |
22 |
11.00% |
Slightly increased |
36 |
18.00% |
No impact |
68 |
34.00% |
Slightly decreased |
47 |
23.50% |
Significantly decreased |
27 |
13.50% |
In addition, the results concerning AI disclosure reveal how transparency affects consumers’ assessments of authenticity. Differences in perceived authenticity under disclosure conditions highlight the role of labeling in shaping initial impressions of AI-generated content. These authenticity-related perceptions constitute an important preliminary stage in the evaluation process, as they precede the formation of more stable beliefs related to credibility and emotional response within the research framework.
3.2.2. Creativity and Attractiveness of AI Ads
The creativity-related results reflect consumers’ evaluations of AI-generated advertising in terms of visual novelty, creative copy, attention capture, and narrative diversity. The distributions reported across Tables 11-14 illustrate how respondents perceive the creative and attractive qualities of AI-generated advertisements at a descriptive level. These findings provide an overview of the extent to which AI-generated advertising is viewed as visually engaging and creatively differentiated from traditional advertising formats.
Table 11. Comparison of novelty and visual impact between AI-generated advertising and traditional advertising.
Sample |
Number of respondents |
Percentage |
Much stronger |
62 |
31.00% |
Slightly stronger |
54 |
27.00% |
About the same |
48 |
24.00% |
Slightly weaker |
23 |
11.50% |
Much weaker |
13 |
6.50% |
Table 12. Consumers’ perceived evaluation of creativity in AI advertising copy.
Sample |
Number of respondents |
Percentage |
Very creative |
38 |
19.00% |
Somewhat creative |
59 |
29.50% |
Neutral |
67 |
33.50% |
Not very creative |
26 |
13.00% |
Not creative at all |
10 |
5.00% |
Table 13. Distribution of consumer preferences for the attractiveness of different types of advertising.
Sample |
Number of respondents |
Percentage |
Live-action advertising |
74 |
37.00% |
Animated or hand-drawn Style advertising |
56 |
28.00% |
AI-generated virtual scenes/characters advertising |
62 |
31.00% |
User-generated content (UGC) advertising |
68 |
34.00% |
Mixed reality advertising (e.g., AR/VR) |
49 |
24.50% |
Respondents were allowed to select multiple options; therefore, percentages exceed 100%.
Table 14. Perceived role of AI technology in enhancing advertising narrative diversity.
Sample |
Number of respondents |
Percentage |
Strongly agree |
41 |
20.50% |
Agree |
52 |
26.00% |
Not sure |
63 |
31.50% |
Disagree |
33 |
16.50% |
Strongly disagree |
11 |
5.50% |
3.2.3. Credibility and Trust Evaluation of AI-Generated Advertising
The results reported in Tables 15-18, as shown in Table 15, summarize consumers’ evaluations of the credibility-related aspects of AI-generated advertising. These tables describe respondents’ perceptions of information reliability, concerns regarding exaggeration or potential manipulation, and the influence of AI disclosure on trust judgments. Together, they capture how consumers assess the trustworthiness of AI-generated advertising content at a descriptive level.
Table 15. Distribution of perceived reliability of AI-generated advertising information.
Sample |
Number of respondents |
Percentage |
More reliable |
62 |
31.00% |
About the same |
57 |
28.50% |
Less reliable |
43 |
21.50% |
Very unreliable |
21 |
10.50% |
Unable to judge |
17 |
8.50% |
Table 16. Level of concern about exaggeration or misleading risks in AI advertising.
Sample |
Number of respondents |
Percentage |
Very concerned |
53 |
26.50% |
Somewhat concerned |
68 |
34.00% |
Neutral |
41 |
20.50% |
Not very concerned |
23 |
11.50% |
Not concerned at all |
15 |
7.50% |
Table 17. Impact of “AI-generated” labeling on information trust.
Sample |
Number of respondents |
Percentage |
Significantly reduced |
39 |
19.50% |
Slightly reduced |
74 |
37.00% |
No impact |
52 |
26.00% |
Slightly increased |
28 |
14.00% |
Significantly increased |
7 |
3.50% |
Table 18. Comparison of trust levels across different advertising sources.
Sample |
Number of respondents |
Percentage |
Brand-official human-created advertising |
88 |
44.00% |
KOL/influencer endorsements |
76 |
38.00% |
Authentic user review videos |
95 |
47.50% |
AI-generated brand promotional advertising |
31 |
15.50% |
News media reports |
63 |
31.50% |
Respondents were allowed to select multiple options; therefore, percentages exceed 100%.
In addition, the comparison of credibility across different information sources illustrates whether trust evaluations are primarily associated with the advertising message itself or with the perceived authority of the brand, platform, or source delivering the content. These credibility-related findings provide an empirical basis for subsequent regression analysis by documenting how perceptions of reliability and risk are distributed across the sample prior to testing the hypothesized relationships.
3.2.4. Emotional Resonance and Brand Connection
The results presented in Tables 19-22 describe consumers’ emotional responses to AI-generated advertising, including the intensity of emotional reactions and perceptions of emotional expression sincerity. These tables document the extent to which AI-generated advertising is perceived as capable of eliciting emotional engagement at a descriptive level.
Table 19. Distribution of perceived emotional resonance.
Sample |
Number of respondents |
Percentage |
Strongly able to resonate |
67 |
33.50% |
Somewhat able to resonate |
58 |
29.00% |
Neutral |
43 |
21.50% |
Not very able to resonate |
24 |
12.00% |
Not able to resonate at all |
8 |
4.00% |
Table 20. Evaluation of the sincerity of emotional expression.
Sample |
Number of respondents |
Percentage |
Very sincere |
39 |
19.50% |
Somewhat sincere |
52 |
26.00% |
Neutral |
61 |
30.50% |
Not very sincere |
36 |
18.00% |
Not sincere at all |
12 |
6.00% |
Table 21. Emotional responses when viewing AI-generated advertising.
Sample |
Number of respondents |
Percentage |
Curiosity |
74 |
37.00% |
Surprise |
62 |
31.00% |
Skepticism |
53 |
26.50% |
Dislike |
38 |
19.00% |
Fondness |
68 |
34.00% |
Indifference |
47 |
23.50% |
Other |
29 |
14.50% |
Respondents were allowed to select multiple options; therefore, percentages exceed 100%.
Table 22. Changes in emotional connection with the brand.
Sample |
Number of respondents |
Percentage |
Yes, significantly enhanced |
27 |
13.50% |
Yes, slightly enhanced |
49 |
24.50% |
No change |
98 |
49.00% |
Decreased instead |
26 |
13.00% |
Overall, the distributions indicate variation in perceived emotional resonance across respondents, reflecting differing evaluations of emotional impact and authenticity in AI-generated content. These findings provide an empirical reference point for subsequent regression analysis by illustrating how emotional responses are distributed prior to examining their mediating role in the relationships between AI-generated advertising features and purchase intention.
The results presented in Tables 23-29 describe respondents’ overall attitudes, perceptions, and behavioral tendencies toward AI-generated advertising. These findings provide a descriptive overview of how consumers evaluate AI-generated advertising in terms of acceptance, perceived development, and its influence on brand-related outcomes. Together, they offer an empirical reference point for subsequent regression analysis by illustrating general response patterns prior to examining the relationships among key variables.
Table 23. Distribution of respondents’ overall attitudes toward AI-generated advertising.
Sample |
Number of respondents |
Percentage |
Strongly support |
62 |
31.00% |
Somewhat support |
58 |
29.00% |
Neutral |
47 |
23.50% |
Somewhat oppose |
24 |
12.00% |
Strongly oppose |
9 |
4.50% |
Table 24. Level of respondents’ recognition of AI advertising as a future trend.
Sample |
Number of respondents |
Percentage |
Strongly agree |
53 |
26.50% |
Mostly agree |
74 |
37.00% |
Not sure |
41 |
20.50% |
Mostly disagree |
22 |
11.00% |
Strongly disagree |
10 |
5.00% |
Table 25. Respondents’ willingness to accept AI-personalized advertising recommendations.
Sample |
Number of respondents |
Percentage |
Very willing |
48 |
24.00% |
Somewhat willing |
56 |
28.00% |
Neutral |
63 |
31.50% |
Not very willing |
26 |
13.00% |
Not willing at all |
7 |
3.50% |
Table 26. Respondents’ evaluation of the technological maturity of AI advertising.
Sample |
Number of respondents |
Percentage |
Very mature |
19 |
9.50% |
Somewhat mature |
38 |
19.00% |
Neutral |
67 |
33.50% |
Not very mature |
52 |
26.00% |
Extremely immature |
24 |
12.00% |
Table 27. Changes in brand modernity or innovative image.
Sample |
Number of respondents |
Percentage |
Significantly improved |
67 |
33.50% |
Slightly enhanced |
58 |
29.00% |
No change |
43 |
21.50% |
Slightly declined |
24 |
12.00% |
Significantly declined |
8 |
4.00% |
Table 28. Impact of AI advertising on brand trust.
Sample |
Number of respondents |
Percentage |
Significantly enhanced |
39 |
19.50% |
Slightly enhanced |
52 |
26.00% |
No impact |
64 |
32.00% |
Slightly weakened |
33 |
16.50% |
Significantly weakened |
12 |
6.00% |
Table 29. Preferences for technologically leading brands.
Sample |
Number of respondents |
Percentage |
Definitely will |
28 |
14.00% |
Probably will |
74 |
37.00% |
Not sure |
55 |
27.50% |
Probably will not |
31 |
15.50% |
Definitely will not |
12 |
6.00% |
3.3. Reliability Assessment and Mediation Examination
Prior to hypothesis testing, reliability analyses were conducted to evaluate the internal consistency of the measurement scales. Cronbach’s alpha coefficients exceeded the recommended threshold of 0.70 for all constructs, indicating satisfactory internal consistency (AI-generated advertising features: α = 0.86; perceived credibility: α = 0.84; emotional resonance: α = 0.88; purchase intention: α = 0.85).
To examine the relationships among AI-generated advertising features, perceived credibility, emotional resonance, and purchase intention, hierarchical multiple regression analyses were conducted. Control variables were entered in Step 1. In Step 2, AI-generated advertising features were included as predictors of perceived credibility and emotional resonance. In Step 3, perceived credibility and emotional resonance were entered simultaneously to assess their effects on purchase intention and to examine potential mediation effects.
The regression results presented in Table 30 indicate that AI-generated advertising features significantly predicted both perceived credibility (β = 0.42, p < 0.001) and emotional resonance (β = 0.47, p < 0.001). In the final model predicting purchase intention, perceived credibility (β = 0.34, p < 0.001) and emotional resonance (β = 0.41, p < 0.001) both demonstrated significant positive effects. The inclusion of these mediating variables led to a significant increase in explained variance (ΔR2 = 0.19, p < 0.001), with the final model explaining 48% of the variance in purchase intention (R2 = 0.48).
These findings suggest that the influence of AI-generated advertising features on purchase intention operates primarily through indirect cognitive and affective mechanisms. Overall, the regression analyses provide empirical support for the proposed relationships among AI-generated advertising features, perceived credibility, emotional resonance, and purchase intention.
Table 30. Hierarchical regression results.
Predictor |
Standardized β |
p-Value |
AI Features → Perceived Credibility |
0.42 |
< 0.001 |
AI Features → Emotional Resonance |
0.47 |
< 0.001 |
Perceived Credibility → Purchase Intention |
0.34 |
< 0.001 |
Emotional Resonance → Purchase Intention |
0.41 |
< 0.001 |
R2 (Final Model) |
0.48 |
|
ΔR2 |
0.19 |
< 0.001 |
Multicollinearity diagnostics were examined prior to interpretation of regression results. Variance Inflation Factor (VIF) values ranged between 1.34 and 2.18, remaining well below the recommended threshold of 5, indicating that multicollinearity was not a concern in the regression models.
3.4. Hypothesis Testing and Generational Differences
Hierarchical regression analyses were conducted to test the proposed hypotheses. The results indicate that AI-generated advertising features positively influence perceived credibility and emotional resonance. In turn, both perceived credibility and emotional resonance positively predict purchase intention, supporting H1 - H4.
Mediation analysis based on regression procedures indicates that the effects of AI-generated advertising features on purchase intention are primarily transmitted through perceived credibility and emotional resonance, supporting H5 and H6.
Effects of Mediators on Purchase Intention
Across generational cohorts, perceived credibility and emotional resonance both demonstrate positive associations with purchase intention. However, the relative strength of these effects varies by generation.
Generation Z.
For Generation Z, emotional resonance (β = 0.46, p < 0.001) demonstrated a stronger effect on purchase intention than perceived credibility (β = 0.28, p < 0.01).
Millennials
Among Millennials, perceived credibility (β = 0.37, p < 0.001) and emotional resonance (β = 0.34, p < 0.001) exerted relatively balanced effects.
Generation X
For Generation X, perceived credibility (β = 0.51, p < 0.001) showed a substantially stronger influence on purchase intention than emotional resonance (β = 0.24, p < 0.05). This finding suggests a credibility-first persuasion mechanism in which trust, transparency, and informational reliability serve as primary determinants of behavioral intention among older consumers.
These findings provide support for H7, indicating that the strength of relationships among variables differs across generational cohorts.
3.5. Comparative Analysis Across Generations
Separate regression analyses conducted for each generational cohort reveal that the effects of AI-generated advertising are not uniform across age groups. Generation Z exhibits stronger associations with emotion- and experience-related factors. Millennials demonstrate relatively balanced effects of credibility and emotional resonance. Generation X places greater emphasis on credibility-oriented evaluations and information reliability.
These generational differences confirm the presence of cohort-based heterogeneity in persuasion mechanisms and highlight that AI-generated advertising effectiveness depends on distinct cognitive and affective pathways across age segments.
4. Discussion
This study examined the mechanisms through which AI-generated advertising influences consumers’ purchase intention, with a particular focus on the mediating roles of perceived credibility and emotional resonance and differences across generational cohorts. The findings provide empirical support for the proposed hypotheses and demonstrate that the persuasive impact of AI-generated advertising is not uniform across consumers; instead, it operates through distinct psychological pathways shaped by generational characteristics.
First, the results confirm that AI-generated advertising features positively influence perceived credibility and emotional resonance, supporting H1 and H2. Furthermore, perceived credibility and emotional resonance both positively influence purchase intention, supporting H3 and H4. The influence of AI-generated advertising on purchase intention is primarily indirect, operating through these two mediating mechanisms rather than through a direct effect. These findings provide support for the mediation hypotheses (H5 and H6). Consistent with persuasion theory, consumer responses to AI-generated advertising involve both cognitive and affective processing. Perceived credibility represents consumers’ rational evaluation of information reliability and authenticity, while emotional resonance captures affective engagement and psychological connection with advertising content. Together, these two pathways form a dual persuasion mechanism through which AI-generated advertising shapes behavioral intention.
Importantly, the findings demonstrate that perceived credibility and emotional resonance function as distinct and complementary mechanisms. High levels of creativity or emotional stimulation alone are insufficient to drive purchase intention if credibility is lacking, while credible but emotionally neutral advertising may fail to sustain attention or engagement. AI-generated advertising therefore achieves stronger persuasive outcomes when both credibility and emotional resonance are simultaneously addressed, highlighting the need for balanced message design.
Second, the study reveals clear generational differences in the dominance of these mediation pathways, supporting H7. For Generation Z, emotional resonance emerges as the primary determinant of purchase intention. This cohort responds more strongly to creative expression, anthropomorphic elements, and emotionally engaging content, reflecting their immersion in interactive digital environments such as short-video platforms and social media. Although credibility remains relevant, emotional engagement constitutes the central route through which AI-generated advertising influences behavior among younger consumers.
For Millennials, the findings indicate a balanced dual-route persuasion pattern. Both perceived credibility and emotional resonance exert comparable influence on purchase intention, suggesting that this cohort integrates affective engagement with rational evaluation. Millennials are receptive to creative and emotionally appealing advertising, yet they also require credible and informative content to support decision-making.
For Generation X, perceived credibility becomes the dominant pathway influencing purchase intention, while emotional resonance plays a secondary role. This pattern suggests a credibility-first persuasion mechanism in which trust, transparency, and information accuracy are critical prerequisites for persuasion.
Taken together, these findings demonstrate that generational cohort acts as an important boundary condition in the persuasion process of AI-generated advertising. Advertising effectiveness is shaped not only by technological features but also by age-related differences in cognitive orientation, media consumption habits, and trust formation processes.
From a theoretical perspective, this study contributes to the literature on AI-driven marketing by integrating generational cohort theory with advertising persuasion mechanisms. It provides empirical support for a dual-path persuasion mechanism of AI advertising effectiveness and highlights generational variation in persuasion processes.
From a practical perspective, the findings suggest that firms should adopt differentiated AI advertising strategies tailored to specific generational segments. For Generation Z, advertisers should emphasize creativity, anthropomorphic design, and emotionally engaging narratives. For Millennials, effective campaigns should integrate credible information with emotional storytelling. For Generation X, AI-generated advertising should prioritize transparency, information reliability, and brand consistency while avoiding excessive novelty.
Overall, this study demonstrates that the effectiveness of AI-generated advertising depends not only on technological sophistication but also on its alignment with consumers’ psychological expectations across generations. AI advertising becomes most persuasive when innovation is balanced with trust and emotional engagement tailored to the characteristics of each generational cohort.
5. Conclusion
This study explored the influence of AI-generated advertising on consumers’ purchase intention by examining the mediating roles of perceived credibility and emotional resonance, as well as differences across generational cohorts. The findings indicate that AI-generated advertising does not produce a uniform persuasive effect across consumers; rather, its effectiveness depends on the specific psychological mechanisms activated and the generational characteristics of the target audience.
The results confirm that perceived credibility and emotional resonance constitute two central mediating pathways through which AI-generated advertising affects purchase intention. Perceived credibility reflects consumers’ cognitive assessment of information reliability and trustworthiness, whereas emotional resonance captures affective engagement and psychological connection with advertising content. The persuasive impact of AI-generated advertising is strongest when these two dimensions are jointly addressed, indicating that neither credibility nor emotional appeal alone is sufficient to fully drive consumer intention.
A key contribution of this study lies in identifying substantial generational differences in the dominance of these mediation mechanisms. For Generation Z, emotional resonance plays a primary role in shaping purchase intention, suggesting that creative design, anthropomorphic elements, and emotionally engaging narratives are particularly effective for younger consumers. For Millennials, perceived credibility and emotional resonance exert relatively balanced effects, reflecting a dual-route persuasion pattern that combines rational evaluation with affective engagement. In contrast, for Generation X, perceived credibility emerges as the dominant determinant of purchase intention, highlighting stronger sensitivity to transparency, risk reduction, and brand familiarity.
From a theoretical perspective, this research extends existing literature on AI-driven marketing by integrating persuasion theory with generational cohort theory in the context of AI-generated advertising. The findings demonstrate that advertising effectiveness is shaped not only by technological features but also by age-related differences in cognitive orientation, trust formation, and media consumption habits. By highlighting generational variation in persuasion mechanisms, this study provides a more nuanced understanding of how intelligent advertising systems influence consumer decision-making.
From a managerial standpoint, the results underscore the importance of adopting differentiated AI advertising strategies across generational segments. Advertisers targeting Generation Z should prioritize emotional storytelling, creativity, and human-like interaction. Campaigns aimed at Millennials should balance credible information with emotional appeal. For Generation X, AI-generated advertising should focus on transparency, informational reliability, and consistency with established brand identities, while avoiding excessive novelty or exaggerated emotional cues.
Despite its contributions, this study has several limitations. First, the use of a cross-sectional survey design limits the ability to infer causal relationships over time. Future research could employ longitudinal or experimental approaches to examine dynamic changes in consumer responses to AI-generated advertising. Second, the sample was restricted to Chinese consumers, which may limit the generalizability of the findings. Cross-cultural studies could further investigate whether the observed generational patterns are consistent across different markets. Finally, while this study focused on perceived credibility and emotional resonance, future research may incorporate additional mediators such as privacy concerns, perceived control, or algorithm transparency to enhance explanatory power.
In conclusion, the effectiveness of AI-generated advertising depends not merely on technological sophistication but on its alignment with consumers’ psychological expectations across generations. By balancing innovation with trust and combining emotional engagement with credibility, AI-generated advertising can serve as an effective and responsible communication tool in the evolving digital marketing landscape.