Consumer Responses to AI-Generated Charitable Giving Ads

This is a self-archived version of an original article. This version may differ from the original in pagination and typographic details. Author(s): Title: Year: Version: Copyright: Rights: Rights url: Please cite the original version: CC BY-NC-ND 4.0 https://creativecommons.org/licenses/by-nc-nd/4.0/ Consumer Responses to AI-Generated Charitable Giving Ads © 2023 the Authors Published version Arango, Luis; Singaraju, Stephen Pragasam; Niininen, Outi Arango, L., Singaraju, S. P., & Niininen, O. (2023). Consumer Responses to AI-Generated Charitable Giving Ads. Journal of Advertising, 52(4), 486-503. https://doi.org/10.1080/00913367.2023.2183285 2023 Full Terms & Conditions of access and use can be found at https://www.tandfonline.com/action/journalInformation?journalCode=ujoa20 Journal of Advertising ISSN: (Print) (Online) Journal homepage: https://www.tandfonline.com/loi/ujoa20 Consumer Responses to AI-Generated Charitable Giving Ads Luis Arango, Stephen Pragasam Singaraju & Outi Niininen To cite this article: Luis Arango, Stephen Pragasam Singaraju & Outi Niininen (2023) Consumer Responses to AI-Generated Charitable Giving Ads, Journal of Advertising, 52:4, 486-503, DOI: 10.1080/00913367.2023.2183285 To link to this article: https://doi.org/10.1080/00913367.2023.2183285 © 2023 The Author(s). Published with license by Taylor & Francis Group, LLC Published online: 08 Mar 2023. Submit your article to this journal Article views: 2819 View related articles View Crossmark data Citing articles: 1 View citing articles Consumer Responses to AI-Generated Charitable Giving Ads Luis Arango a , Stephen Pragasam Singaraju b , and Outi Niininen c a University of Queensland, St. Lucia, Queensland, Australia; b Universiti Teknologi Brunei, Bandar Seri Begawan, Brunei Darussalam; c Jyv€ askyl€ a University, Jyv€ askyl€ a, Finland ABSTRACT Content created by employing artificial intelligence (AI) algorithms, also known as synthetic content, promises to radically change the advertising and marketing landscape in the coming decades, presumably for the better. It is fundamental for advertising and marketing scholars and practitioners to have solid knowledge of how synthetic content is perceived by consumers before widespread adoption is promoted. Across three experimental studies we tested how consumers in charitable giving contexts reacted to advertising messages featuring content generated by an AI neural network. We show that potential donors responded differently to children’s faces when they knew they had been generated by AI. Study 1 established that awareness of the falsity of a face or its status as an AI-generated image has a negative impact on donation intentions. This negative impact is serially mediated by empathy and anticipatory guilt and empathy and emotion perception. Study 2 investigated several motives for employing AI-generated images and indicated that charities employing those images can benefit by making their ethical motives salient. Finally, Study 3 revealed that under extraordinary circumstances the use of AI images by charities is considered acceptable by consumers and is likely to lead to similar outcomes as the use of real images. Therefore, we recommend a cautious approach to the adoption of synthetic content. Recent studies (Campbell et al. 2021; Whittaker et al. 2020) indicate that advertising messages will transition to an era dominated by artificial intelligence (AI) in the coming decades. Drivers of this evolutionary process are cost and convenience, among others. A company called Generated Photos offers a glimpse into what the future might hold for advertising content creation. As stated on the company’s website, for a small fee an organization can purchase a humanlike AI-generated model that is indistinguishable from a real human model by consumers. Organizations can reuse the model as many times as they would like and hold exclusive rights to it, saving significant money on professional photographers, models, and the many aspects involved in photo shoots (e.g., makeup, logistics). As the example of Generated Photos indicates, it is difficult to discern what would dissuade companies from extensively adopting AI-generated content, particularly because of the return on investment (ROI) this advertising strategy promises. However, before rushing into AI-generated content adoption, organizations need to consider the effects of such content on consumers. This article examines the effects of AI-generated (also known as synthetic) ads (Whittaker et al. 2020)oncharity advertising, which can be defined as media messages delivered through mass media channels that seek to promote charities’goals (Grau 2014). We focus on charities for two reasons. First, advertising, as a percentage of the organization’s budget, tends to be higher for profit-oriented companies compared to charities (The NonProfit Times 2016). Charity advertising budgets are generally small, and these organizations consequently might be especially motivated to employ affordable AI-generated content. Second, as the effective altruism movement (MacAskill 2015) and the giving multiplier initiative (Caviola and Greene 2022) show, societies embrace the prospect of widespread effectiveness when it comes to charity work (i.e., we all want charities to spend their CONTACT Luis Arango [email protected] The University of Queensland Business School, 39 Blair Dr, St Lucia, QLD 4067, Australia. Luis Arango (Master of Marketing, La Trobe University) is a doctoral student, The University of Queensland Business School. Stephen Singaraju (PhD, Monash University) is an assistant professor, UTB School of Business, Universiti Teknologi Brunei. Outi Niininen (PhD, University of Surrey) is a marketing program coordinator, Jyv€ askyl€ a University School of Business and Economics. ß2023 The Author(s). Published with license by Taylor & Francis Group, LLC This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License (http://creativecommons.org/licenses/by-ncnd/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited, and is not altered, transformed, or built upon in any way. The terms on which this article has been published allow the posting of the Accepted Manuscript in a repository by the author(s) or with their consent. JOURNAL OF ADVERTISING 2023, VOL. 52, NO. 4, 486–503 https://doi.org/10.1080/00913367.2023.2183285 resources as effectively as possible). If, for instance, AIgenerated content is more (or less) effective at enticing donors than real content, then we want to establish that. Given the central role of emotional appeals in charity advertising (Septianto and Tjiptono 2019), the question of whether synthetic ads are useful as a tool to promote charities’work is framed here as an inquiry into the effects of this type of advertising on consumer emotions that precede donation intentions. In other words, we aimed to understand consumers’emotional reactions when presented with these ads and their consequences for effective charity appeals. We did so by investigating how awareness of the falsity of synthetic content, when used on charitable giving advertising messages, impacts empathy and—via the empathic response—guilt and emotion perception. Even though research on emotion has flourished for several decades in many disciplines, beginning in the 1960s with the pioneering work of Paul Ekman and Silvan Tomkins (Ekman, Sorenson, and Friesen 1969; Pollak, Camras, and Cole 2019; Tomkins 1962), a recent call for further research on advertising identified emotions as a promising route (Poels and Dewitte 2019). Furthermore, a review of the charity advertising literature recommends more studies focused on structural relations between variables, especially considering the high frequency of factorial designs in this area (Wymer and Gross 2021). The present research sits at the intersection of two advertising research fields: AI and consumer emotions. In addition to contributing to the literature in understudied areas, this study contributes to marketing and advertising research and practice in at least three other ways. First, this is one of the first studies to empirically study consumer reactions to synthetic content (Sands et al. 2022) and to define a new subdomain in consumer behavior research. Thus, this study is expected to be one of the first to open a new and underexplored area of consumer perceptions of synthetic content, contributing to the vitality of the advertising and marketing fields (MacInnis 2011). Synthetic content is likely to play a dominant role in the future of advertising and marketing, but little is known about how consumers perceive it, in part due to its novelty (Campbell et al. 2021; Sands et al. 2022). Second, we uncover some of the psychological mechanisms behind consumer responses to synthetic content by employing stimuli (faces) that are pervasive in marketing and advertising communications. Finally, we study boundary conditions and reveal some circumstances that are likely to positively affect perceptions of synthetic content. Based on these findings, we offer strategies that can be employed when using this type of content. This article proceeds as follows. First, we discuss synthetic content and the motivation for our research by identifying the lacuna in the current advertising knowledge it intends to fill. Second, we develop our conceptual model based on several theoretical approaches to the phenomena of awareness of falsity, empathy, guilt, and emotion perception (Basil, Ridgway, and Basil 2008; Campbell et al. 2021). Later, in Study 1, we empirically test this model using stimuli generated by a neural network. After this, we introduce two additional studies motivated by the findings of Study 1. Finally, we discuss our results, note the limitations of our work, and suggest avenues for future research. Literature Review Artificial Intelligence and Content Creation AI-generated media, also known as synthetic or generative media, is a type of content poised to revolutionize advertising and marketing in the coming years (Campbell et al. 2021). Synthetic media is the output of AI algorithms and represents a new stage in the evolution of content creation. Synthetic content can be classified into several categories based on diverse criteria, the most popular being deepfakes—synthetic media that result from the replacement of aspects of a media source by an AI algorithm which have the potential to deceive. A popular deepfake showcases a scene from The Shining with the face of Jim Carrey superimposed onto Jack Nicholson’s character (https://www.youtube. com/watch?v=HG_NZpkttXE). By contrast, novel synthetic content can be exemplified by AI-generated faces of nonexistent individuals, as presented on the website https://thispersondoesnotexist.com/ (see Figure 1). These images are fake in the sense that they do not correspond to the image of any real person, but due to their shocking realism, they have the potential to Figure 1. AI-generated face (left) versus digitally created face (right). JOURNAL OF ADVERTISING 487 induce the belief that they do correspond to a real person. The revolutionary character of synthetic content is partly explained by the manipulation possibilities offered by AI algorithms compared to digital or analog techniques. The difference is illustrated in Figure 1,wherethe face on the left is generated using digital technology and the one on the right is generated by the technology employed in this article—an AI neural network known as a generative adversarial network (GAN). Even though the digital image is realistic, it is still distinguishable from a real one, in contrast to the AI-generated image. Synthetic media has an edge over digital media in at least two other respects. The first is accessibility, or the democratization of content creation (Karnouskos 2020). Synthetic media is usually open source; people can freely access code and modify it according to their specific needs. Furthermore, training algorithms to create synthetic content does not require access to immense amounts of computing power, so it is affordable for individuals. The second is personalization and data integration. Through the application of big data, AI has enabled the practice of mass personalization of communication content, which results in better persuasion outcomes (Hermann 2021;Singarajuetal. 2022). In the case of charity advertising, algorithms can collect social media data and modify or replace images in real time to make a call to action more enticing (e.g., present potential donors with images that data suggest stand abetterchanceofengagingthem). Despite its great potential, empirical research on synthetic media is still in its infancy. A bibliographic analysis of the AI literature in business disciplines (Feng et al. 2021) points to the novelty of synthetic media or content as one reason for its limited study. Deepfakes, for example, only gained popularity in 2018, mainly as entertainment devices on social media platforms. Only a few conceptual models and corresponding research agendas have been suggested to guide the study of synthetic content in marketing and advertising (Campbell et al. 2021,2022; Whittaker, Letheren, and Mulcahy 2021), yet empirical work is still missing. A pivotal goal of this article is to address the need for empirical research in this nascent field by studying consumer responses to synthetic content. We focus on consumer attitudes, emotions, and intentions triggered by hyperrealistic synthetic faces featured in ad templates designed to promote charities’causes. Some research has been done on consumer reactions to AI in general, mostly when it is employed as a tool to influence, inform, and predict behavior through data-mining techniques (Davenport et al. 2020). Studies in this area have unveiled phenomena such as so-called algorithm aversion (Dietvorst, Simmons, and Massey 2015)—or people’s supposedly negative bias toward interacting with algorithms in certain settings (Jussupow, Benbasat, and Heinzl 2020). However, even if this research is illuminating, it is not clear that it can account for the peculiarities that could be present in scenarios where content is created by AI or where users interact with synthetic media like deepfakes. For instance, Epstein et al. (2020) explored people’s perceptions of AI agency after a portrait titled Edmond the Belamy, which was marketed as the first painting created by an AI algorithm, sold for $432,500. The financial success of the painting can hardly be explained by invoking algorithm aversion. To the best of our knowledge, to date, the only studies that have researched consumer responses to synthetic content have been carried out by Thomas and Fowler (2021) and Sands et al. (2022), who investigated people’s responses to AI influencers of the likes of Lil Miquela. Some of their findings, which also run counter to the phenomenon of algorithm aversion, suggest that AI influencers are viewed as positively as real ones across several dimensions. Nonetheless, even if highly interesting, a significant limitation of the research by Thomas and Fowler (2021) is the fact that they used vignettes, and consumers were not exposed to the AI influencers themselves but rather were asked to imagine scenarios featuring them. By contrast, Sands et al. (2022) used synthetic content. In line with this, our study presents people with synthetic content. Awareness of Falsity This article follows the idea that a fundamental element of consumers’responses to synthetic images is awareness of falsity, which is interpreted as presented reality (Campbell et al. 2021). Consumers, in general, display negative attitudes toward falsity in ads (Held and Germelmann 2018). Perceived falsity leads to negative reactions by consumers, such as a defensive approach when processing information and distrust in communications, which constitute barriers to persuasive attempts (Boush, Friestad, and Wright 2015; Darke and Ritchie 2007). However, for this article, it is important to decouple awareness of falsity and manipulation intent. In for-profit environments, these elements frequently commingle. Companies that, for instance, exaggerate the benefits of using a product do so intentionally, knowing that their portrayal of the product is not accurate but is motivated by financial gain. Research has shown that consumer inferences of 488 L. ARANGO ET AL. manipulative intent frequently thwart advertisers’ goals (Cotte, Coulter, and Moore 2005). With synthetic content, it is crucial to realize that falsity might not lead to inferences of manipulative intent, especially when there is a reasonable expectation that the consumer will be able to detect the untrue elements of an ad. For example, “Malaria Must Die,”a 2019 campaign led by a team of scientists, doctors, and activists, features a video of David Beckham speaking nine languages, including Hindi and Mandarin (https:// www.youtube.com/watch?v=QiiSAvKJIHo). Even though it is not true that David Beckham speaks these languages, consumers are likely to know this. Given this expectation (and other elements, such as the fact that the intention of “Malaria Must Die”is not to deceive but to connect with its audience), consumers can perceive the falsity of the ad and simultaneously withhold any judgment regarding manipulative intent on the part of the advertiser. We can distinguish at least two non–mutually exclusive routes that consumers might follow to ascertain the falsity of synthetic content. Given hyperrealistic content, consumers can rely on their background knowledge and infer that the content is not authentic. This route follows a top-down information-processing path because, in this case, perception is informed by personal factors such as knowledge (Gregory 1970; Pieters and Wedel 2004). We can call this the inferential route to falsity awareness. Conversely, consumers can base their falsity assessments on inherent features of the content, processing information in a bottom-up fashion with the stimulus itself determining their response (e.g., Gibson 2014; Pieters and Wedel 2004). We call this the direct route to falsity awareness. As technology progresses and synthetic content increasingly improves, consumers will likely come to rely more on the inferential route or technologies currently under development. Thanks to many efforts, such as the Deepfake Detection Challenge by Kaggle, a technological route to identifying falsity will eventually open. This study separated manipulative intent and awareness falsity. Subjects who were informed about the falsity of an image (i.e., its status as a piece of AIgenerated content) were also asked to imagine a scenario in which a charity was open about the fake nature of the image. This approach has the obvious benefit of differentiating between two different constructs, but it is also intended as a way of adding external validity to the study. If charities were to use these images, they would not likely do so deceptively (hiding their nature as AI generated). This would risk people’s positive perceptions of charity organizations (Noble and Wixley 2014). However, dissociating falsity and manipulative intent forces us to consider alternative ways in which falsity might affect consumers. In the next section, we begin building a conceptual model, arguing that falsity affects empathy. Empathy Empathy is an equivocal term; it can be construed as simply sharing an emotion with another agent (Hoffman 1985) or as the emotion of sympathy (Batson 2014), among others. Here, we focus on empathy as a cognitive skill, particularly one that allows individuals to mentally occupy the place of others, or as the capacity of individuals to take up the first-person perspective of another agent. This is an advanced cognitive process, as it implies the suppression of an egocentric perspective, also known as perspective taking or cognitive empathy (Davis 2015). Importantly, the evaluation of empathy in this study was episodic and not dispositional; we assessed empathy as a situational reaction and not as a personality trait. Empathy is a cognitive process susceptible to psychological distance. We tend to empathize more with close others than with distant others, with closeness being a function of many factors, such as spatiotemporal distance (Liberman, Trope, and Stephan 2007). This is why people find it extremely difficult to care about the environmental impacts of climate change on future humans (Pahl and Bauer 2013) or the plight of individuals in remote locations (Joseph and Xin 2012). Identifiability plays a role in psychological distance phenomena, with studies showing, for instance, that individuals are more willing to help children if they have seen pictures of them compared to a scenario where they have to rely exclusively on a description of their situation (Small and Loewenstein 2003). Presumably, identifiability facilitates the simulation process through which the individual imagines the target’s disadvantaged position (Chambers and Davis 2012). For this study, psychological distance was a crucial trait of empathy as a cognitive phenomenon. Throughout this study, the main stimuli were charity ads displaying faces generated by AI technology. These faces did not correspond to real persons despite their hyperrealism. We believed that once subjects were made aware of the falsity of AI-generated images, psychological distance would be created, leading to a reduced empathic response. Following the previous discussion, we felt that individuals would find it challenging to identify with the potential recipients of their help, as the AI-generated images they were to be presented with did not correspond to any of JOURNAL OF ADVERTISING 489 them. We held that this effect would be present even if subjects knew that the images were representative of the situation or had features of real donation recipients and knew that a charity employing such images did not act with manipulative intent. Therefore, we hypothesized the following: H1: Awareness of falsity (i.e., of synthetic images) will negatively impact empathy. A large body of evidence links prosocial behavior and empathy conceptualized as situational perspective taking. While empathy as perspective taking is not invariably linked to prosocial behavior (e.g., Caruso, Epley, and Bazerman 2006;Galinskyetal.2008), several studies have found a positive relationship between experiencing empathy and prosociality. For instance, studies have linked instructions to imagine the perspective of others to helping behaviors directed at them (e.g., Batson et al. 1989;Stocks,Lishner,andDecker2009), including marginalized group members (Aberson and Haag 2007)or members of out-groups (Galinsky and Moskowitz 2000). A reduced empathic reaction in consumers as a result of awareness of the falsity of AI-generated images was expected to have negative consequences in terms of donation intentions (and subsequent behaviors). Here, we tested two possible mechanisms through which reduced empathy could harm donation intentions. First, we argue that reduced empathy results in weaker emotional reactions to ads, focusing our attention on guilt. Second, we claim that reduced empathy lowers the perceived emotional intensity of the faces displayed in ads. Both of these routes will lead to lower donation intentions. Empathy and Guilt Empathy has sometimes been shown to precede socalled moral emotions (Silfver et al. 2008).Inanowclassical work, Haidt (2003) defined moral emotions as those that “respond to moral violations or that motivate moral behavior”(p. 853). Moral emotions can be grouped into two categories: other-condemning and self-conscious (Haidt 2003). Other-condemning emotions, such as anger or contempt, are directed at parties that are perceived as displaying unethical behavior, such as a business mistreating its employees or contaminating the environment, and can lead to punishing behaviors, such as negative word of mouth (Grappi, Romani, and Bagozzi 2013). Self-conscious emotions, such as guilt or shame, are directed at the self, and they typically constitute responses to instances of personal (in)action that are perceived negatively, particularly in the sense that they do not comply with the ethical or moral standards embraced by the individual. Here, we focus on guilt, mainly because of the ubiquitous character and effectiveness of guilt appeals in charitable donation settings (Basil, Ridgway, and Basil 2006; Hibbert et al. 2007; Urbonavicius et al. 2019). The type of guilt discussed in this article is anticipatory: it is guilt that the subject anticipates feeling upon dismissing charitable giving ads and their calls to action. In other words, in this scenario, the main motivation for the subject to display positive attitudes toward or engagement in donation behavior would be to avoid the aversive consequences (feelings of guilt) of not doing so. Anticipatory guilt as a response to charity advertising has been found to be preceded by empathy (Basil, Ridgway, and Basil 2008). In this study, we also construed empathy as an antecedent of guilt and put forth two related hypotheses. First, we hypothesized that there is a positive relationship between empathy and guilt; that is, subjects who are better able to imagine themselves in the position of potential recipients of their help will experience more anticipatory guilt. In addition, following Basil, Ridgway, and Basil (2008), we believed that due to the unpleasantness of guilt, a positive relationship would be found between anticipatory guilt and donation intentions. Therefore, the more guilt an individual anticipates feeling, the more positive his or her attitude toward donating will be. Formally, we offer these two hypotheses: H2: Empathy will positively impact anticipatory guilt. H3: Anticipatory guilt will positively impact donation intentions. We also expected awareness of falsity to affect donation intentions via the empathy–anticipatory guilt route (i.e., we expected to find a negative relationship between awareness of falsity and donation intentions serially mediated by empathy and guilt). Individuals aware of the falsity of AI images would then be less empathic, which would lead to less anticipatory guilt and lower donation intentions. H4: The negative impact of awareness of falsity on donation intentions is serially mediated by empathy, then anticipatory guilt. Empathy and Emotion Perception Emotion perception can be defined as the “perception of emotion expressed by another person verbally, facially, with the body, or through a combination of these modalities”(Olderbak and Wilhelm 2017,p.1093).Although emotions are usually expressed multimodally, we focus on the facial expressions of emotions and the corresponding capacity of consumers to perceive emotions expressed in this way. This focus can be justified on 490 L. ARANGO ET AL. several grounds. Charitable organizations’emotional appeals almost invariably use visual stimuli as a main component, such as human faces, especially of children (Cao and Jia 2017). Such extended practices by advertising and marketing practitioners align with research stating that humans are highly visual creatures (Kaas and Balaram 2014) who have developed the ability to extract large amounts of information from facial expressions (Tsao and Livingstone 2008). Emotion perception is a fundamental process in charitable giving research (Tong et al. 2021). Many studies (Bagozzi and Moore 1994; Genevsky and Knutson 2015; Small and Verrochi 2009; ZemackRugar and Klucarova-Travani 2018) have attempted to establish what type of facial expression (e.g., sad versus happy) is more effective at encouraging donations, or at least effective at creating positive attitudes toward them, such as increasing donation intentions. Although findings in this particular respect are mixed, they demonstrate the effectiveness of portraying faces displaying emotions in charity ads as a tool to promote giving. In this study, we employed faces displaying emotions with a negative valence: sadness. According to psychological models that specify the relationship between empathy and emotion perception (Mayer, Caruso, and Salovey 1999; Orchard et al. 2009), perceiving emotions in others is an effortful and not automatic process preceded by, among other emotions, empathy. In these models, emotion perceivers are portrayed as first going through a cognitive process whereby they occupy the perspectives of the agents expressing the emotion and empathizing with them. These models indicate a positive relationship between cognitive empathy and emotion perception. This implies that hindering or blocking individuals’capacity to exercise cognitive empathy has a deleterious effect on their emotion perception skills. As previously argued, we believe that AI-generated images constitute a barrier to individuals trying to empathize with advertising messages. Individuals who are presented with an ad portraying an AI-generated image and who are informed about the falsity of the image are subsequently expected to show a decreased capacity to perceive the emotion expressed by the face presented. Given the previous discussion, we hypothesized the following: H5: Empathy will positively impact emotion perception. H6: Emotion perception will positively impact donation intentions. Finally, we also expected awareness of falsity to affect donation intentions via the alternative empathy–emotion perception route (i.e., we expected to find a negative relationship between awareness of falsity and donation intentions serially mediated by empathy and emotion perception). Individuals who are aware of the falsity of an image would then be less empathic, which would lead to a decreased capacity to perceive the emotion of the face portrayed in the ad and lower donation intentions. H7: The negative impact of awareness of falsity on donation intentions is serially mediated by empathy, then emotion perception. The conceptual model in Figure 2 depicts the hypothesized relationships (hypotheses 4 and 7 are the top and bottom paths, respectively, from awareness of falsity to donation intentions). Methods Ethics and Data Management This study was preregistered at the Credibility Lab (https://credlab.wharton.upenn.edu/). We did not collect any personally identifiable information; therefore, this study does not fall under the General Data Protection Regulation (GDPR), which is one of the European Union’s data privacy regulations (except for a pretest). An advanced ethical assessment of the study was conducted, and this research complies with the guidelines of the Finnish National Board on Research Integrity TENK. Code, supplementary statistical outputs, preregistration, and materials are accessible on the Open Science Framework platform at https://osf.io/ fn9h4/. Participant recruitment was done through CloudResearch (Litman, Robinson, and Abberbock 2017), except for the first survey and pretest, which used Prolific panels (Palan and Schitter 2018). AI Images The images employed in this study are the product of a GAN trained on a data set of human faces. GANs are machine-learning algorithms that comprise two deep neural networks, a generator, and a discriminator that (through a competitive training process) create novel outputs. In general, the generator first creates a fake image of a person. This image and one from the data set serve as inputs for the discriminator, whose task is to distinguish one from the other (discriminate between the real and the fake image produced by the generator). The outcome of the process, or whether the discriminator was successful at telling the images apart, serves as feedback for subsequent iterations of the process. The generator increasingly improves its ability to produce fake images (images that are more difficult for JOURNAL OF ADVERTISING 491 the discriminator to differentiate from real images), while the discriminator improves its ability to determine which is which. Images generated by the generator can then be used to create nonexistent images (in our case, the faces of nonexistent children). There are three considerations when using GAN images. First, the images are novel, as GAN algorithms are different from others that are exclusively discriminative, such as facial recognition algorithms. Second, this novelty ensures that our study complied with the GDPR and the Biometric Information Privacy Act (BIPA), as GAN-generated images do not share the biometric properties of faces belonging to particular, real people. Third, we used a StyleGAN (Karras, Laine, and Aila 2019); there are several GAN algorithms (e.g., the Mean and Covariance Feature Matching GAN [McGan] and the Maximum Mean Discrepancy GAN [MMD GAN]) that utilize, among others, different objective functions to measure the distance between the generated and original data distributions. StyleGAN is a state-of-the-art technique that avoids several shortcomings of previous GAN algorithms and is thus ideal for generating novel outputs. Preliminary Survey and Images Pretest To ascertain the relevance of the present study, we carried out a short survey of attitudes toward AI images, employing a sample of individuals involved in charity work. The general attitude toward the images was positive. Those against the use of the images saw them as lacking “soul”and were worried about their potential to connect with donors. Those in favor thought the images would be effective because no real children would be exposed, which would protect their privacy and dignity (Steeves 2006). In addition, the subjects were not able to differentiate between AI-generated and real images. Therefore, consumers cannot follow what we previously called direct or indirect routes to establish content falsity. This result replicates previous findings (K€ obis, Dole zalov a, and Soraperra 2021). Study 1 Sample The sample size was determined a priori (Soper 2021), per the following parameters: statistical power: 0.8; minimum anticipated effect size: 0.2; number of constructs: 5; number of observed variables: 12; and p value: 0.05. The recommended sample size for a structural model with those parameters was 376, and we successfully collected 464 responses. (The survey completion rate for the aware group was 43.3%; for the not-aware group, it was 69.7%.) After multivariate outliers were excluded, the remaining 458 participants were divided into two groups. The not-aware group consisted of 234 subjects (47.9% female), and the aware group comprised 224 subjects (50% female). Procedures To manipulate awareness of falsity, we divided the subjects into two groups: The participants in the aware group were presented with a charity ad that included an AI-generated image, and they were informed that the image was generated by AI technology (see Figure 3). The technology was briefly explained, and the subjects were asked to think of a charity using the image in a nonmanipulative fashion (i.e., not hiding the status of the image as AI generated). The participants in the notaware group were not informed about the status of the image as AI generated. Filtering questions were employed in the questionnaire to test the subjects’comprehension of the technology. Several versions of the ad in Figure 3 were created corresponding to the different Figure 2. Conceptual model. 492 L. ARANGO ET AL. For charities not working on disaster relief, even if synthetic content can offer several benefits (e.g., savings in terms of time and budget), the benefits must be carefully weighed against the negative impact that such content can have on important outcomes, such as donation intentions or charity reputation. In these scenarios, charities are advised to wait to adopt synthetic content. Before adopting the trend of synthetic content, charities should closely follow any potential changes in public attitudes toward AI technology (Vasiljeva, Kreituss, and Lulle 2021), introduce it slowly (if at all), and monitor consumer data to ascertain its effectiveness. As we have shown, content authenticity is strongly advised at this time. Future Research Ideally, practitioners’use of synthetic content should be guided and informed by the strong theoretical foundations developed in academia. Otherwise, as our research shows, synthetic content use can potentially inflict damage on organizations. Researchers should extend our work and help define the conditions under which synthetic content use is a safe venture for organizations. Due to its novelty, the field of consumer reactions to synthetic content offers an oversupply of research possibilities. The scope of the present research was the not-forprofit context of charity organizations. Although our conceptual model is not expected to be applicable in for-profit settings, some of its constructs and the theoretical paradigm on which it is based might be relevant to those contexts. Here, we list some valuable research questions that scholars could explore in this respect. First, we found that manipulative intent was significantly higher for individuals who knew they were being presented with AI images. A for-profit motive could worsen this, with consumers aware of the synthetic nature of content reacting defensively toward it (Cotte, Coulter, and Moore 2005; Darke and Ritchie 2007). Second, research suggests that psychological distance is related not only to empathy but also to constructs such as trust, with information sources that are perceived as closer being more trusted than those who are not (Sands et al. 2022). Could the use of synthetic content in marketing and advertising communications become a barrier to achieving a trusting relationship between consumers and companies? A case in point is Synthesia (https://www.synthesia.io/), a platform that offers companies the ability to create videos featuring AI avatars, “saving up to 80% of their time and budget.”An inescapable question is whether companies can expect a good ROI from the use of such avatars to promote their products, despite the time and budget savings in marketing and advertising campaigns. AI avatars could significantly hurt pivotal outcomes, such as consumer trust, and increase perceptions of manipulative intent. In addition, the use of AI avatars or images could significantly decrease the effectiveness of emotional appeals, as our findings on emotion perception suggest. An appeal might not be as effective if consumers know that the person they see smiling in an ad is not a real person but a Synthesia avatar. These are empirical questions worth exploring. There are two additional avenues for future research that are relevant to both the not-for-profit and for-profit contexts. The first is moderation relationships. Consumers with certain personality traits, holding certain beliefs, or belonging to certain demographic groups could react differently to the inherent falsity of synthetic ads. Determining which consumers are likelier to react negatively to synthetic ads can protect organizations against the misuse of synthetic ads. Second, longitudinal studies can be undertaken to ascertain how public attitudes toward synthetic ads change as general adoption grows. The phenomenon of adoption by marketing and advertising practitioners and by the general public, which can be approached using models such as the technology acceptance model and its subsequent extensions (Davis 1989; Venkatesh et al. 2003), can eventually normalize synthetic content and lead to consumer indifference toward its falsity. Establishing whether the artificial character of synthetic content might eventually become irrelevant for consumers—and, if so, when—can help conservative organizations better time their transition to such content. Limitations We would like to note the limitations of this research. The first is that we used images exclusively. Even though charities use images of children on a regular basis, other media, particularly videos, are an important part of their campaigns. A GAN could also be employed to superimpose fake faces; however, creating novel, credible videos resembling those employed by charities is still beyond the capabilities of these technologies. Second, even though we went to great lengths to ensure data quality by following appropriate recommendations (Aguinis, Villamor, and Ramani 2021), and data collected online is, for the most part, reliable (Kees et al. 2017), it is important that future studies attempt a replication of results with different samples. Finally, even though we followed previous studies (Ahmed 2021; Sands et al. 2022) and manipulated awareness of falsity using JOURNAL OF ADVERTISING 499 disclosure labels, consumer skepticism regarding the fake nature of the images might remain due to their hyperrealism. Such skepticism could be particularly marked among consumer segments not familiar with new technologies. Future studies could employ more detailed explanations of AI technology and the remarkably real content outputs it can produce. Conflict of Interest Statement The authors declare that there are no conflicts of interest. Funding Data collection for this study was funded by Jyv€ askyl€ a University School of Business and Economics. ORCID Luis Arango http://orcid.org/0000-0002-1581-8732 References Aberson, C. L., and S. C. Haag. 2007.“Contact, Perspective Taking, and Anxiety as Predictors of Stereotype Endorsement, Explicit Attitudes, and Implicit Attitudes.” Group Processes & Intergroup Relations 10 (2):179–201. doi:10.1177/1368430207074726. Aguinis,H.,I.Villamor,andR.S.Ramani.2021.“MTurk Research: Review and Recommendations.”Journal of Management 47 (4):823–37. doi:10.1177/0149206320969787. Ahmed, S. 2021.“Fooled by the Fakes: Cognitive Differences in Perceived Claim Accuracy and Sharing Intention of Non-Political Deepfakes.”Personality and Individual Differences 182:111074. doi:10.1016/j.paid.2021.111074. Amengual, M., and E. P. Apfelbaum. 2021.“True Motives: Prosocial and Instrumental Justifications for Behavioral Change in Organizations.”Management Science 67 (8): 5032–51. doi:10.1287/mnsc.2020.3708. Bagozzi, R. P., and D. J. Moore. 1994.“Public Service Advertisements: Emotions and Empathy Guide Prosocial Behavior.”Journal of Marketing 58 (1):56. doi:10.2307/ 1252251. Basil,D.Z.,N.M.Ridgway,andM.D.Basil.2006.“Guilt Appeals: The Mediating Effect of Responsibility.”Psychology &Marketing23 (12):1035–54. doi:10.1002/mar.20145. Basil,D.Z.,N.M.Ridgway,andM.D.Basil.2008.“Guilt and Giving: A Process Model of Empathy and Efficacy.” Psychology & Marketing 25 (1):1–23. doi:10.1002/mar.20200. Batson, C. D. 2014.The Altruism Question: Toward a Social-Psychological Answer. New York: Psychology Press. Batson, C. D., J. G. Batson, C. A. Griffitt, S. Barrientos, J. R. Brandt, P. Sprengelmeyer, and M. J. Bayly. 1989. “Negative-State Relief and the Empathy—Altruism Hypothesis.”Journal of Personality and Social Psychology 56 (6):922. doi:10.1037/0022-3514.56.6.922 Boush, D. M., M. Friestad, and P. Wright. 2015.Deception in the Marketplace: The Psychology of Deceptive Persuasion and Consumer Self-Protection. New York: Routledge. Campbell, C., K. Plangger, S. Sands, and J. Kietzmann. 2021. “Preparing for an Era of Deepfakes and AI-Generated Ads: A Framework for Understanding Responses to Manipulated Advertising.”Journal of Advertising 51: 22–38. doi:10.1080/00913367.2021.1909515. Campbell, C., K. Plangger, S. Sands, J. Kietzmann, and K. Bates. 2022.“How Deepfakes and Artificial Intelligence Could Reshape the Advertising Industry: The Coming Reality of AI Fakes and Their Potential Impact on Consumer Behavior.”Journal of Advertising Research 62 (3):241–51. doi:10.2501/JAR-2022-017. Campbell,M.C.1995.“When Attention-Getting Advertising Tactics Elicit Consumer Inferences of Manipulative Intent: The Importance of Balancing Benefits and Investments.” Journal of Consumer Psychology 4(3):225–54. doi:10.1207/ s15327663jcp0403_02. Cao, X., and L. Jia. 2017.“The Effects of the Facial Expression of Beneficiaries in Charity Appeals and Psychological Involvement on Donation Intentions: Evidence from an Online Experiment.”Nonprofit Management & Leadership 27 (4):457–73. doi:10.1002/nml.21261. Caruso, E., N. Epley, and M. Bazerman. 2006.“When Perspective Taking Increases Taking: Reactive Egoism in Social Interaction.”Journal of Personality and Social Psychology 91 (5):872–89. Caviola, L., and J. Greene. 2022.“GivingMultipler.” Accessed January 16, 2022. https://givingmultiplier.org/. Chambers, J. R., and M. H. Davis. 2012.“The Role of the Self in Perspective-Taking and Empathy: Ease of Self-Silmulation as a Heuristic for Inferring Empathic Feelings.”Social Cognition 30 (2):153–80. doi:10.1521/soco.2012.30.2.153. Cotte, J., R. A. Coulter, and M. Moore. 2005.“Enhancing or Disrupting Guilt: The Role of Ad Credibility and Perceived Manipulative Intent.”Journal of Business Research 58 (3): 361–8. doi:10.1016/S0148-2963(03)00102-4. Darke, P. R., and R. J. B. Ritchie. 2007.“The Defensive Consumer: Advertising Deception, Defensive Processing, and Distrust.”Journal of Marketing Research 44 (1):114– 27. doi:10.1509/jmkr.44.1.114. Davenport, T., A. Guha, D. Grewal, and T. Bressgott. 2020. “How Artificial Intelligence Will Change the Future of Marketing.”Journal of the Academy of Marketing Science 48 (1):24–42. doi:10.1007/s11747-019-00696-0. Davis, F. D. 1989.“Perceived Usefulness, Perceived Ease of Use, and User Acceptance of Information Technology.” MIS Quarterly 13 (3):319–40. doi:10.2307/249008. Davis, M. H. 1980.“A Multidimensional Approach to Individual Differences in Empathy.”JSAS Catalog of Selected Documents in Psychology 10 (85):2–19. Davis, M. H. 2015.“Empathy and Prosocial Behavior.”In The Oxford Handbook of Prosocial Behavior, edited by D. Schroeder and W. Graziano, 282–306. Oxford/New York: Oxford University Press. Dietvorst,B.J.,J.P.Simmons,andC.Massey.2015. “Algorithm Aversion: People Erroneously Avoid Algorithms after Seeing Them Err.”Journal of Experimental Psychology. General 144 (1):114–26. doi:10.1037/xge0000033 500 L. ARANGO ET AL. Ekman, P., E. R. Sorenson, and W. V. Friesen. 1969.“PanCultural Elements in Facial Displays of Emotion.” Science 164 (3875):86–8. doi:10.1126/science.164.3875.86 Ellemers, N., J. van der Toorn, Y. Paunov, and T. van Leeuwen. 2019.“The Psychology of Morality: A Review and Analysis of Empirical Studies Published from 1940 through 2017.”Personality and Social Psychology Review 23 (4):332–66. doi:10.1177/1088868318811759. Epstein, Z., S. Levine, D. G. Rand, and I. Rahwan. 2020. “Who Gets Credit for AI-Generated Art?”iScience 23 (9):101515. doi:10.1016/j.isci.2020.101515. Feng, C. M., A. Park, L. Pitt, J. Kietzmann, and G. Northey. 2021.“Artificial Intelligence in Marketing: A Bibliographic Perspective.”Australasian Marketing Journal. 29 (3):252–63. doi:10.1016/j.ausmj.2020.07.006. Galinsky, A. D., W. W. Maddux, D. Gilin, and J. B. White. 2008.“Why It Pays to Get inside the Head of Your Opponent: The Differential Effects of Perspective Taking and Empathy in Negotiations.”Psychological Science 19 (4):378–84. doi:10.1111/j.1467-9280.2008.02096.x. Galinsky, A. D., and G. B. Moskowitz. 2000.“PerspectiveTaking: Decreasing Stereotype Expression, Stereotype Accessibility, and in-Group Favoritism.”Journal of Personality and Social Psychology 78 (4):708–24. doi:10. 1037//0022-3514.78.4.708 Garrido, M. V., and M. Prada. 2017.“KDEF-PT: Valence, Emotional Intensity, Familiarity and Attractiveness Ratings of Angry, Neutral, and Happy Faces.”Frontiers in Psychology 8:2181. doi:10.3389/fpsyg.2017.02181. Genevsky, A., and B. Knutson. 2015.“Neural Affective Mechanisms Predict Market-Level Microlending.” Psychological Science 26 (9):1411–22. doi:10.1177/ 0956797615588467. Gibson, J. J. 2014.The Ecological Approach to Visual Perception: Classic Edition. New York: Psychology Press. Goldsmith, R. E., B. A. Lafferty, and S. J. Newell. 2000. “The Impact of Corporate Credibility and Celebrity Credibility on Consumer Reaction to Advertisements and Brands.”Journal of Advertising 29 (3):43–54. doi:10. 1080/00913367.2000.10673616. GPT3. 2020, September 8. “A Robot Wrote This Entire Article. Are You Scared Yet, Human?”The Guardian. https://www.theguardian.com/commentisfree/2020/sep/ 08/robot-wrote-this-article-gpt-3. Grappi, S., S. Romani, and R. P. Bagozzi. 2013.“Consumer Response to Corporate Irresponsible Behavior: Moral Emotions and Virtues.”Journal of Business Research 66 (10):1814–21. doi:10.1016/j.jbusres.2013.02.002. Grau, S. L. 2014.Marketing for Nonprofit Organizations: Insights and Innovations. Oxford: Oxford University Press. Gregory, R. L. 1970.The Intelligent Eye. London: Weidenfeld &Nicolson. Haidt, J. 2003.“The Moral Emotions.”In Handbook of Affective Sciences, edited by R. J. Davidson, K. R. Scherer, and H. H. Goldsmith, 852–70. New York: Oxford University Press. Hayes, A. F. 2018.Introduction to Mediation, Moderation, and Conditional Process Analysis: A Regression-Based Approach. 2nd ed. New York: The Guilford Press. Held,J.,andC.C.Germelmann.2018.“DeceptioninConsumer Behavior Research: A Literature Review on Objective and Perceived Deception.”Projectics/Proy ectica/Projectique 3(21): 119–45. Hermann, E. 2021.“Artificial Intelligence and Mass Personalization of Communication Content—An Ethical and Literacy Perspective.”New Media & Society 24 (5): 1258–77. doi:10.1177/14614448211022702. Hibbert, S., A. Smith, A. Davies, and F. Ireland. 2007. “Guilt Appeals: Persuasion Knowledge and Charitable Giving.”Psychology & Marketing 24 (8):723–42. doi:10. 1002/mar.20181. Hock, S. J., R. Bagchi, and T. M. Anderson. 2020. “Promotional Games Increase Consumer Conversion Rates and Spending.”The Journal of Consumer Research 47 (1):79–99. doi:10.1093/jcr/ucz043. Hoffman, M. L. 1985.“Interaction of Affect and Cognition in Empathy.”In Emotions, Cognition, and Behavior, edited by C. E. Izard, J. Kagan, and R. B. Zajonc, 103–31. New York: Cambridge University Press. James, V. 2021, December 6. “This AI Art App Is a Glimpse at the Future of Synthetic Media.”The Verge. https://www.theverge.com/2021/12/6/22820106/ai-artapp-dream-synthetic-media-wombo. Joseph, P., and H. Xin. 2012.“Empathy, Donation, and the Moderating Role of Psychological Distance.”Advances in Consumer Research 40:250. Jussupow, E., I. Benbasat, and A. Heinzl. 2020.“Why Are We Averse towards Algorithms? A Comprehensive Literature Review on Algorithm Aversion.”European Conference on Information Systems Proceedings. https:// aisel.aisnet.org/ecis2020_rp/168. Kaas, J. H., and P. Balaram. 2014.“Current Research on the Organization and Function of the Visual System in Primates.”Eye and Brain 6 (Suppl):1–4. doi:10.2147/EB. S64016. Kamal, Y. 2016.“Study of Trend in Digital Marketing and Evolution of Digital Marketing Strategies.”International Journal of Engineering Science 6 (5):5300–2. Karnouskos, S. 2020.“Artificial Intelligence in Digital Media: The Era of Deepfakes.”IEEE Transactions on Technology and Society 1 (3):138–47. doi:10.1109/TTS. 2020.3001312 Karras, T., S. Laine, and T. Aila. 2019.“A Style-Based Generator Architecture for Generative Adversarial Networks.”Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, Long Beach, CA. Kees, J., C. Berry, S. Burton, and K. Sheehan. 2017.“An Analysis of Data Quality: Professional Panels, Student Subject Pools, and Amazon’s Mechanical Turk.”Journal of Advertising 46 (1):141–55. doi:10.1080/00913367.2016. 1269304. Kim, N. 2014.“Advertising Strategies for Charities: Promoting Consumers’Donation of Time versus Money.”International Journal of Advertising 33 (4):707– 24. doi:10.2501/IJA-33-4-707-724. K€ obis,N.C.,B.Dole  zalov a, and I. Soraperra. 2021.“Fooled Twice: People Cannot Detect Deepfakes but Think They Can.”iScience 24 (11):103364. doi:10.1016/j.isci.2021.103364. Liberman, N., Y. Trope, and E. Stephan. 2007. “Psychological Distance.”In Social Psychology: Handbook of Basic Principles, edited by A. Kruglanski and T. Higgins, Vol. 2, 353–83. New York: Guilford Press. JOURNAL OF ADVERTISING 501 Litman, L., J. Robinson, and T. Abberbock. 2017. “TurkPrime.com: A Versatile Crowdsourcing Data Acquisition Platform for the Behavioral Sciences.” Behavior Research Methods 49 (2):433–42. doi:10.3758/ s13428-016-0727-z. MacAskill, W. 2015.Doing Good Better: Effective Altruism and a Radical New Way to Make a Difference. New York: Guardian Faber Publishing. MacInnis, D. J. 2011.“A Framework for Conceptual Contributions in Marketing.”Journal of Marketing 75 (4):136–54. doi:10.1509/jmkg.75.4.136. Mayer, J. D., D. R. Caruso, and P. Salovey. 1999. “Emotional Intelligence Meets Traditional Standards for an Intelligence.”Intelligence (Norwood) 27 (4):267–98. doi:10.1016/S0160-2896(99)00016-1. Noble, J., and S. Wixley. 2014.“Matter of Trust: What the Public Thinks of Charities and How It Affects Trust.” London: NPC. http://www.thinknpc.org/publications/ matter-of-trust. Olderbak, S., and O. Wilhelm. 2017.“Emotion Perception and Empathy: An Individual Differences Test of Relations.”Emotion 17 (7):1092–106. doi:10.1037/ emo0000308. Orchard, B., C. MacCann, R. Schulze, G. Matthews, M. Zeidner, and R. D. Roberts. 2009.“New Directions and Alternative Approaches to the Measurement of Emotional Intelligence.”In Assessing Emotional Intelligence: Theory, Research, and Applications, edited by C. Stough, D. H. Saklofske, & J. D. A. Parker, 321–44. New York: Springer. doi:10.1007/978-0-387-88370-0_17. Osborne, J., and A. Overbay. 2004.“The Power of Outliers (and Why Researchers Should Always Check for Them).” Practical Assessment, Research, and Evaluation 9 (1):1–8. Pahl, S., and J. Bauer. 2013.“Overcoming the Distance: Perspective Taking with Future Humans Improves Environmental Engagement.”Environment and Behavior 45 (2):155–69. doi:10.1177/0013916511417618. Palan, S., and C. Schitter. 2018.“Prolific.ac—A Subject Pool for Online Experiments.”Journal of Behavioral and Experimental Finance 17:22–7. doi:10.1016/j.jbef.2017.12. 004. Pieters, R., and M. Wedel. 2004.“Attention Capture and Transfer in Advertising: Brand, Pictorial, and Text-Size Effects.”Journal of Marketing 68 (2):36–50. doi:10.1509/ jmkg.68.2.36.27794. Poels, K., and S. Dewitte. 2019.“The Role of Emotions in Advertising: A Call to Action.”Journal of Advertising 48 (1):81–90. doi:10.1080/00913367.2019.1579688. Pollak, S. D., L. A. Camras, and P. M. Cole. 2019.“Progress in Understanding the Emergence of Human Emotion.” Developmental Psychology 55 (9):1801. Qin, X., and Z. Jiang. 2019.“The Impact of AI on the Advertising Process: The Chinese Experience.”Journal of Advertising 48 (4):338–46. doi:10.1080/00913367.2019. 1652122. Rogers, E. M. 2010.Diffusion of Innovations. New York: Simon and Schuster. Rønnow-Rasmussen, T. 2015.“Intrinsic and Extrinsic Value.”In The Oxford Handbook of Value Theory, edited by I. Hirose and J. Olson, 29–43. Oxford: Oxford University Press. Sands, S., C. L. Campbell, K. Plangger, and C. Ferraro. 2022.“Unreal Influence: Leveraging AI in Influencer Marketing.”European Journal of Marketing 56 (6):1721– 47. doi:10.1108/EJM-12-2019-0949. Septianto, F., and F. Tjiptono. 2019.“The Interactive Effect of Emotional Appeals and past Performance of a Charity on the Effectiveness of Charitable Advertising.”Journal of Retailing and Consumer Services 50:189–98. doi:10. 1016/j.jretconser.2019.05.013. Short, A. 2021, May 27. “13 AI Tools That Are at the Cutting Edge of Music Production.”Happy.https://happymag.tv/ai-music-production-tools/. Silfver,M.,K.Helkama,J.-E.L € onnqvist, and M. Verkasalo. 2008.“The Relation between Value Priorities and Proneness to Guilt, Shame, and Empathy.”Motivation and Emotion 32 (2):69–80. doi:10.1007/s11031-008-9084-2. Singaraju, S. P., J. L. Rose, L. A. Arango-Soler, C. D’Souza, S. M. S. Khaksar, and A. R. Brouwer. 2022.“The Dark Age of Advertising: An Examination of Perceptual Factors Affecting Advertising Avoidance in the Context of Mobile Youtube.” Journal of Electronic Commerce Research 23 (1):13–32. Small, D. A., and G. Loewenstein. 2003.“Helping a Victim or Helping the Victim: Altruism and Identifiability.” Journal of Risk and Uncertainty 26 (1):5–16. doi:10.1023/ A:1022299422219. Small, D. A., and N. M. Verrochi. 2009.“The Face of Need: Facial Emotion Expression on Charity Advertisements.” Journal of Marketing Research 46 (6):777–87. doi:10. 1509/jmkr.46.6.777. Soper, D. 2021.“A-Priori Sample Size Calculator for Structural Equation Models.”https://www.danielsoper. com/statcalc. Spears, N., and S. N. Singh. 2004.“Measuring Attitude toward the Brand and Purchase Intentions.”Journal of Current Issues & Research in Advertising 26 (2):53–66. doi:10.1080/10641734.2004.10505164. Steeves, V. 2006.“It’s Not Child’s Play: The Online Invasion of Children’s Privacy.”University of Ottawa Law & Technology Journal 3 (1):169–88. Stocks, E. L., D. A. Lishner, and S. K. Decker. 2009. “Altruism or Psychological Escape: Why Does Empathy Promote Prosocial Behavior?”European Journal of Social Psychology 39 (5):649–65. doi:10.1002/ejsp.561. The NonProfit Times.2016, April 20. “Charities Spent 4c on Advertising for Each $1 Raised Online.”The NonProfit Times.https://www.thenonprofittimes.com/npt_articles/ 1-digital-donors-cost-charities-4%C2%A2/#::text=Non profits%20are%20spending%20an%20average,generation% 20and%20new%20donor%20acquisition%20.%20%20https:// www.webstrategiesinc.com/blog/how-much-budget-foronline-marketing-in-2014. Thomas, V. L., and K. Fowler. 2021.“Close Encounters of the AI Kind: Use of AI Influencers as Brand Endorsers.” Journal of Advertising 50 (1):11–25. doi:10.1080/ 00913367.2020.1810595. Tomkins, S. 1962.Affect Imagery Consciousness: Volume I: The Positive Affects. New York: Springer Publishing Company. Tong, Z., M. Yi, W. Feng, Y. Yu, D. Liu, and J. Zhang. 2021.“The Interaction of Facial Expression and DonorRecipient Eye Contact in Donation Intentions: Based on 502 L. ARANGO ET AL. the Intensity of Emotion.”Frontiers in Psychology 12: 661851. doi:10.3389/fpsyg.2021.661851. Tsao, D. Y., and M. S. Livingstone. 2008.“Mechanisms of Face Perception.”Annual Review of Neuroscience 31 (1): 411–37. doi:10.1146/annurev.neuro.30.051606.094238. Urbonavicius, S., K. Adomaviciute, I. Urbutyte, and J. Cherian. 2019.“Donation to Charity and Purchase of Cause-Related Products: The Influence of Existential Guilt and Experience.”Journal of Consumer Behaviour 18 (2):89–96. doi:10.1002/cb.1749. Vasiljeva, T., I. Kreituss, and I. Lulle. 2021.“Artificial Intelligence: The Attitude of the Public and Representatives of Various Industries.”Journal of Risk and Financial Management 14 (8):339. doi:10.3390/ jrfm14080339. Venkatesh, V., M. G. Morris, G. B. Davis, and F. D. Davis. 2003.“User Acceptance of Information Technology: Toward a Unified View.”MIS Quarterly 27 (3):425–78. doi:10.2307/30036540. Whittaker, L., T. C. Kietzmann, J. Kietzmann, and A. Dabirian. 2020.““All Around Me Are Synthetic Faces”: The Mad World of AI-Generated Media.”IT Professional 22 (5):90–9. doi:10.1109/MITP.2020.2985492. Whittaker, L., K. Letheren, and R. Mulcahy. 2021.“The Rise of Deepfakes: A Conceptual Framework and Research Agenda for Marketing.”Australasian Marketing Journal 29 (3):204–14. doi:10.1177/1839334921999479. Wood, C., M. Conner, E. Miles, T. Sandberg, N. Taylor, G. Godin, and P. Sheeran. 2016.“The Impact of Asking Intention or Self-Prediction Questions on Subsequent Behavior: A Meta-Analysis.”Personality and Social Psychology Review 20 (3):245–68. doi:10.1177/ 1088868315592334. Wymer, W., and H. Gross. 2021.“Charity Advertising: A Literature Review and Research Agenda.”Journal of Philanthropy and Marketing,1–21. doi:10.1002/nvsm. 1723. Zagefka, H., and T. James. 2015.“The Psychology of Charitable Donations to Disaster Victims and beyond: The Psychology of Charitable Donations.”Social Issues and Policy Review 9 (1):155–92. doi:10.1111/sipr.12013. Zemack-Rugar, Y., and S. Klucarova-Travani. 2018.“Should Donation Ads Include Happy Victim Images? The Moderating Role of Regulatory Focus.”Marketing Letters 29 (4):421–34. doi:10.1007/s11002-018-9471-8. JOURNAL OF ADVERTISING 503