Unreal influence: leveraging AI in influencer marketing

Abstract

Purpose This paper aims to examine how consumers respond to social media influencers that are created through artificial intelligence (AI) and compares effects to traditional (human) influencers. Design/methodology/approach Across two empirical studies, the authors examine the efficacy of AI social media influencers. With Study 1, the authors establish baseline effects for AI influencers and investigate how social-psychological distance impacts consumer perceptions. The authors also investigate the role of an influencer’s agency – being autonomous or externally managed – to test the boundaries of the results and determine the interactive effects between influencer type and influencer agency. Study 2 acts as an extension and validation of Study 1, whereby the authors provide generalisability and overlay the role of need for uniqueness as a moderated mediator. Findings The authors show that there are similarities and differences in the ways in which consumers view AI and human influencers. Importantly, the authors find no difference in terms of intention to follow or personalisation. This suggests that consumers are equally open to follow an AI or human influencer, and they perceive the level of personalisation provided by either influencer type as similar. Furthermore, while an AI influencer is generally perceived as having lower source trust, they are more likely to evoke word-of-mouth intentions. In understanding these effects, the authors show that social distance mediates the relationship between influencer type and the outcomes the authors investigate. Results also show that AI influencers can have a greater effect on consumers who have a high need for uniqueness. Finally, the authors find that a lack of influencer agency has a detrimental effect. Research limitations/implications The studies investigate consumers’ general response to AI influencers within the context of Instagram, however, future research might examine consumers’ response to posts promoting specific products across a variety of category contexts and within different social media platforms. Practical implications The authors find that in some ways, an AI influencer can be as effective as a human influencer. Indeed, the authors suggest that there may be a spill-over effect from consumer experiences with other AI recommendation systems, meaning that consumers are open to AI influencer recommendations. However, the authors find consistent evidence that AI influencers are trusted less than traditional influencers, hence the authors caution brands from rushing to replace human influencers with their AI counterparts. Originality/value This paper offers novel insight into the increasingly prominent phenomenon of the AI influencer. Specifically, it takes initial steps towards developing understanding as to how consumers respond to AI influencers and contrast these effects with human influencers.

Buffer bots: The role of virtual service agents in mitigating negative effects when service fails

“…A second stream of research concerns the consumer reactions to virtual agents in service encounters (i.e., Chung et al, 2020; Robinson et al, 2020; Sands et al, 2021; Verhagen et al, 2014). Like their human counterparts, virtual service agents can influence customer service evaluations and behavior (Cassell et al, 2000; Sands, Ferraro, et al, 2022; Sands, Campbell, Plangger, et al, 2022). For instance, interactions with virtual service agents can influence customer satisfaction when the interaction meets or exceeds expectations (Santini et al, 2018), subsequently influencing purchase intentions (Chung et al, 2020; Wiedmann et al, 2009).…”

The three-step persuasion model on YouTube: A grounded theory study on persuasion in the protein supplements industry

“…Arguably, such virtual influencers can fulfill a user's social needs just as good as any human could do (Arsenyan and Mirowska, 2021 ). However, findings indicate that (fashion) AI influencers are often perceived as less trustworthy (Sands et al, 2022 ).…”

“…Research has started to explore the use of AI to rank multimedia influencers based on audience and content representation to match them with marketeers (Farseev et al, 2018(Farseev et al, , 2021. In addition, the effect of AI and virtual SMIs on consumers has been investigated in different studies (Arsenyan and Mirowska, 2021;Block and Lovegrove, 2021;Sands et al, 2022). Arguably, such virtual influencers can fulfill a user's social needs just as good as any human could do (Arsenyan and Mirowska, 2021).…”

What causes social media users to engage and mimic virtual influencers? The role of self-congruity

“…Furthermore, the contents of SMIs can be powerful because of their characteristics, such as attractiveness, expertise, authenticity, and trustworthiness (Ki & Kim, 2019;Wiedmann & von Mettenheim, 2021), making them excellent message sources in marketing communication. More recently, the practice of SMIs has introduced a new form of influencers, virtual or artificial intelligence influencers (AI influencers) (Sands et al, 2022a), which are referred to as virtual influencers (VIs).…”