How AI encourages consumers to share their secrets? The role of anthropomorphism, personalisation, and privacy concerns and avenues for future research

Purpose – This paper aims to explore the overall research question “ How can arti ﬁ cial intelligence (AI) in ﬂ uence consumer information disclosure? ” . It considers how anthropomorphism of AI, personalisation and privacy concerns in ﬂ uence consumers ’ attitudes and encourage disclosure of their private information. Design/methodology/approach – This research draws upon the personalisation-privacy paradox (PPP) and privacy calculus theory (PCT) to address the research question and examine how AI can in ﬂ uence consumer information disclosure. It is proposed that anthropomorphism of AI and personalisation positively in ﬂ uence consumer attitudes and intentions to disclose personal information to a digital assistant, while privacy concerns negatively affect attitude and information disclosure. Findings – This paper develops a conceptual model based on and presents seven research propositions (RPs) forfuture research. Originality/value – Building upon PPP and PCT, this paper presents a view on the bene ﬁ ts and drawbacks of AI from a consumer perspective. This paper contributes to literature by critically re ﬂ ecting upon on the question how consumer information disclosure is in ﬂ uenced by AI. In addition, © Bianca Kronemann, Hatice Kizgin, Nripendra Rana and Yogesh K. Dwivedi. Published in Spanish Journal of Marketing – ESIC . Published by Emerald Publishing Limited. This article is published under the Creative Commons Attribution (CC BY 4.0) licence. Anyone may reproduce, distribute, translate and create derivative works of this article (for both commercial and non-commercial purposes), subject to full attribution to the original publication and authors. The full terms of this licence maybe


Introduction
Artificial intelligence (AI) is contributing to a shift towards a more algorithmic society (Shankar, 2018).The exchange of value in the digital environment is changing as technology is changing from being a tool that is used by actors to being an actor in the value exchange itself (Novak and Hoffman, 2019).Human-less transaction mediated by intelligent technology is growing in numbers and frequency (Hofacker and Corsaro, 2020).Advancements in AI are now also being used across several consumer industries such as retail and e-commerce as well as in marketing.One popular application of AI in marketing is in form conversational agents (Rai, 2020;Thomaz et al., 2020).These are based on natural language computer programmes designed to approximate human speech and interact with people via a digital interface (Thomaz et al., 2020).
However, AI may not be welcomed favourably by all stakeholders due to the circumstance that AI implies the increased ability of technology owners to collect, analyse, combine and control consumer data (Mazurek and Malagocka, 2019).The interaction with digital assistants leaves a rich digital footprint of personal and behavioural data, which means that consumers become active participants in the erosion of their own privacy (Pomfret et al., 2020).Thus, the topic of privacy is under heightened attention across academic researchers, social critics as well as regulators (Martin and Murphy, 2019) because in our knowledge-based digital market environment, information privacy and security have become the primary consumer concerns related to the internet, algorithms and data (Brill et al., 2019;Hong and Thong, 2013;Kietzmann et al., 2018;Mazurek and Malagocka, 2019).The notion of data privacy is critical because advancements in AI are based on and have accelerated with the availability of large amounts of data and computational capabilities to analyse them.The free flow of data is critical because data is the fuel of AI success (Mazurek and Malagocka, 2019).
Protecting consumer privacy and user information is critical to the expanding integration of digital assistants into business models as well as everyday life.Businesses face a substantial risk of abandoned investment into digital assistants if consumers lack trust in either the technology or the company to protect their privacy and personal information (Pappas, 2016).Existing research finds that privacy concerns can form a barrier to consumer adoption of new technology.However, recent research suggests that consumers showcase a trade-off between benefits and risks (Vilmakumar et al., 2021).Based on this ambiguity in the existing literature, this paper will critically reflect on the topics of privacy concerns in relation to AI, benefits in terms of personalisation offered by AI, and consumer information disclosure to AI to answer the overall research question of "How can AI influence consumer information disclosure?".
This paper contributes to the growing discussion in the marketing literature around consumers' interaction with AI, benefits of personalisation versus drawbacks of privacy concerns, and consumer information disclosure.To contribute to existing literature, this study will draw upon the personalisation-privacy paradox (PPP) to examine the effects of AI on consumer attitudes and consumer information disclosure.The PPP is a valuable theoretical basis because it describes a situation in which consumers appreciate the value of personalisation offered by a technology (Aguirre et al., 2015;Bleier and Eisenbeiss, 2015) while being aware of marketers' exploitation of their personal data which raises privacy concerns (Cloarec, 2020).This idea is also reflected in privacy calculus theory (PCT) (Dinev and Hart, 2006;Wang et al., 2016) which is also drawn upon to better understand how and why consumers assess benefits and drawbacks of AI and decide whether to disclose their private information.Based on PCT and PPP, this research proposes a conceptual model with seven research propositions (RPs) that address how anthropomorphism of AI, personalisation and privacy concerns affect consumers' attitudes towards AI and their intention to disclose personal information to AI.

Role of anthro pomorphism
The remaining parts of this paper are structured into four sections: Firstly, literature relating to consumer interaction with AI is reviewed to provide background information on the increased need to better understand consumer interaction with AI, as well as consequences of such interaction.Secondly, the theoretical foundation for a research model is presented, and eight RPs introduced.This section will cover the topics of anthropomorphism of AI, privacy concerns in relation to AI, personalised interactions with AI, and the PPP.Thirdly, the paper will discuss insights and findings gained from the review of literature and development of a research model.Finally, a conclusion is drawn that included reflection on limitations and suggestions for future research.
2. Literature review: consumer interaction with AI Interaction between AI and consumers is becoming a central topic of discussion in the marketing and service literature.Companies adopt these conversational agents as new frontfacing customer service that interact with consumers during service encounters (Lariviere et al., 2017;Thomaz et al., 2020;van Doorn et al., 2017;Wilson and Daugherty, 2018).It is suggested that AI will fundamentally change the nature of interaction between companies and their customers (Haenlein and Kaplan, 2019;Ramaswamy and Ozcan, 2018).
However, research relating to the impact of AI is still developing and relatively sparse (Davenport et al., 2020;Steinhoff et al., 2019).Table 1 gives an overview of selected studies, organised by publication date, that examine AI in the marketing, service and hospitality context.The table highlights each paper's theoretical foundation and research aim, key findings of each paper as well as limitations.
From reviewing above selected studies, four key themes emerge, which will be covered in the following paragraphs: adoption and use of AI technology; anthropomorphism of AI; benefits of AI; and potential barriers to AI success such as privacy and security concerns.
The overview of previous studies shows that most of the previous literature has focused on consumer use and adoption of AI, highlighting that AI integration and application in the consumer context is not a fad but a rising trend (Gursoy et al., 2019).Research shows that consumers are rapidly embracing AI for daily use because they enable individuals to access timely and useful information and can meet customer demand for contextually relevant and highly personalised content that is delivered in real-time (Brill et al., 2019;Canbek and Mutlu, 2016).In addition, research finds that the ease of using AI and its ability to provide relevant information are key factors in influencing consumers to use them, which also is found to influence loyalty (Moriuchi, 2019).Further studies find that a variety of factors along social, relational and functional dimensions are important for adoption of AI (Fernandes and Oliveira, 2021;Vilmakumar et al., 2021).Research shows that individuals are mainly motivated to adopt AI based on functional elements, but they also find support for social and relational elements that drive adoption (Fernandes and Oliveira, 2021).In addition, experience with technology is important to consider in the development of attitudes and adoption intentions of new technology like AI.
Anthropomorphism of AI is a second main theme across literature.The underlying idea for anthropomorphism is that it can influence consumers to think about intelligent agents more favourably.The positive influence of increased human likeness is also confirmed in a study by Belanche et al. (2021) who found that human likeness positively affects utilitarian, social, monetary and emotional value expectations.However, another study examines the uncanny valley theory and find support for their hypotheses that too humanlike features lead to negative evaluations due to perceived threats to human identity (Kim et al., 2019).Yet it remains unclear perceptions of uncanniness translate into behaviour.This is in line with another study that found anthropomorphism as an impeding determinant of consumers'

Role of anthro pomorphism
acceptance of service robots due to the reason that they pose a threat to users' identity (Lu et al., 2019).AI offers many benefits to both companies in terms of improved capabilities to collect, analyse and use data in real-time, which creates competitiveness but also the opportunity to co-create value with consumers through interaction.Yet, the paper does not consider consumer attitudes around the availability and provision for that data that is necessary for real-time co-creation.Companies like Amazon or Apple are exemplary for the utilisation of AI, either sold separately to consumers (e.g.Amazon's Echo) or included in existing consumer devices (e.g.Siri).As an operating system of iOS, Siri is one of the most salient applications of Apple that has caught attention of a tech-savvy audience (Canbek and Mutlu, 2016).One of the most significant advantages of these assistants is their conversational interface (Vilmakumar et al., 2021) and the opportunity to meet the demand for contextually relevant and personalized information that is delivered to the consumer in real-time (Brill et al., 2019).Furthermore, AI has the advantage of being highly scalable, and the ability to deliver routine customer service to large numbers of people simultaneously (Chui et al., 2018;Davenport et al., 2020;Duan et al., 2019;Wilson and Daugherty, 2018).
The increased ability of AI to collect, analyse and store consumer data, however, also results in concerns relating to privacy and security of data.Privacy concerns are relevant to consider for adoption because they also influence trust (Vilmakumar et al., 2021).Existing literature shows that privacy concerns have a negative impact on intentions to use a technology, engagement and attitude towards an innovative technology (Brill et al., 2019;Grewal et al., 2021).Consumers have also privacy concerns because they do not trust AI.

SJME
Overall, it becomes clear that perceived risk of AI poses challenges that need to be investigated to understand implications for AI.
Above discussed studies shed first light on the adoption, potential benefits and barriers as well as consequences of the integration of AI.However, it becomes evident that most studies have been from a use and adoption perspective.There is a lack of critical reflection around the necessity of data and personal information that is required for the successful delivery of AI in the first place.It has been neglected to fully examine consumer concerns and potential barriers in relation to data privacy and information disclosure to AI.

Theoretical background and research propositions
The following section will present the theoretical background of this study being PCT and the PPP and highlight how drawing upon these theories will offer new insights on how consumers may be encouraged to disclose their private information to AI.In addition, this section will propose a research model with seven RPs.

Personalisation-privacy paradox and privacy calculus theory
In Western developed societies, individuals have increasingly raised privacy and data security concerns (Foehr and Germelmann, 2020).Becoming increasingly aware "that they are being watched", consumers are concerned about their data being misused or abused due to eroding confidence that firms are respecting data privacy (Fitzgerald, 2019).Individuals are challenged to manage the complex trade-offs that new technological innovation brings.It must be evaluated how the benefits of a new technology compare to the risks of information privacy concerns (Acquisti et al., 2015) This trade-off has been highlighted as the PPP (Aguirre et al., 2015;Bleier and Eisenbeiss, 2015) which describes the process of consumers¨critical assessment of offered benefits and added value of personalisation against marketers¨exploitation of their private information which raises privacy concerns (Cloarec, 2020).However, the evaluation can be difficult for consumers due to the circumstance that technology often functions as a black box with a lack of evidence of where consumer data is stored, or who will be able to access it (Hofacker and Corsaro, 2020).
Reflecting the idea of the PPP, and to gain a better understanding of privacy and disclosure behaviour, the concept of privacy calculus or PCT will be drawn upon in addition to the PPP.According to the privacy calculus, consumers trade off their privacy or risks to their privacy in a cognitive process of rationality against the prospect of receiving benefit or utility (Dinev and Hart, 2006;Wang et al., 2016).Benefits that personalisation offers include improved experience through curation of relevant content or discounts (Zhu et al., 2017).Central to the privacy calculus perspective is the underlying assumption that consumers are empowered to exercise control over their private information (Dinev et al., 2016), and that rational consumers are apprehensive towards revealing personal information.PCT has been adopted in previous research to gain a better understanding of how consumers evaluate information disclosure (Gutierrez et al., 2019;Pentina et al., 2016;Sun et al., 2015;Xu et al., 2011).For the context of this research, this means, how consumers evaluate benefits of anthropomorphised interaction and personalisation versus privacy risks and the question how this assessment influences their attitudes and information disclosure intentions.PPP and privacy calculus offer therefore a useful theoretical foundation to examine how consumers evaluate the benefits of personalisation against the risks of privacy concerns and its implications on consumer information disclosure.

Proposed conceptual model and research propositions
Based on the discussion of PCT and the concept of anthropomorphism, a conceptual research model is proposed that examines consumer interaction with AI-based digital assistants and their information disclosure towards AI-based digital assistants in the light of privacy concerns.The conceptual model is depicted in Figure 1 below.The model builds upon PCT and anthropomorphism to explain consumer intentions to disclose information and includes eight RPs.In the following paragraphs, the RPs will be discussed and presented.
3.2.1 Anthropomorphism of artificial intelligence.Anthropomorphism has received considerable attention in the marketing and consumer behaviour literature because it represents an opportunity for marketers to affect consumption or affect consumer behaviour related to consumption (Epley, 2018).Anthropomorphism occurs when humans perceive other things as like themselves (Hume, 1757(Hume, /1957) ) or "imbue real or imagined behaviour of nonhuman agents with humanlike characteristics, motivations, intentions or emotions" (Epley et al., 2007, p. 864) which will influence how humans react to those objects.The concept of anthropomorphism is relevant in the context of AI because companies try to humanise their technology by assigning them humanoid traits, e.g.looks, name and voice, which has the aim to encourage customers to converse and bond with them (Novak and Hoffman, 2019;Schweitzer et al., 2019;Steinhoff et al., 2019;van Doorn et al., 2017).Across academic literature, there is a debate to whether anthropomorphism positively affects consumer attitudes, intentions and use of technology, or whether is serves as impeding variable.For instance, previous studies find that anthropomorphism has negative effects on consumer attitudes towards AI and is an impeding determinant of consumers¨acceptance of service robots due to the reason that they pose a threat to users' identity (Gursoy et al., 2019;Lu et al., 2019).Here, reference to the uncanny valley is made (Mori, 1970) which describes the effects of how more humanlike features lead to positive evaluations, until too humanlike features result in eeriness and anxiety in response to the technology.In contrast to that, it has been suggested that anthropomorphism can improve levels of engagement, collaboration and interaction (Steinhoff et al., 2019;Yang et al., 2022).
For this study, it is assumed that anthropomorphism of AI technology has the potential to nudge consumers to greater self-disclosure of data due to the circumstance that it will transmit social cues to consumers that activate social scripts (Thomaz et al., 2020).Previous research has proven that consumers treat computers like social actors if they display a minimum of social cues, such as asking questions or sharing information with the consumer and that consumers feel obliged to respond with also sharing information due to learned social scripts (Nass and Moon, 2000).Based on this and the insights derived from existing Anthropomorphism is a critical construct to better understand human interactions with technology (van Doorn et al., 2017).The authors point out that a product with a smiley face is perceived as friendlier than a product that looks techy.To make AI appear friendlier, designers can assign anthropomorphic appearances, and even create an identity for AIsystems (Ene et al., 2019).Moreover, anthropomorphism can influence consumers to perceive AI as more vivid and potentially treat AI like a sentient being, which has been proven to increase psychological warmth and positively affects consumers attitudes (Kim et al., 2019).Thus, it is proposed that: RP2. Anthropomorphism of AI positively affects consumers' attitudes towards the digital assistant.

Privacy concerns.
Privacy is defined as "the ability of the individual to control the terms under which personal information is acquired and used" (Westin, 1967, p. 7) and privacy in the online environment as "an individual's control and awareness of the collection and usage or personal data" (Hann et al., 2007).In the marketing context, discussions around information disclosure and privacy have in the past been influenced by an understanding of privacy as personal responsibility, which requires individuals to exercise control over the collection, use and distribution of their personal data (Pomfret et al., 2020), what contrasts with a philosophical approach to data privacy and data protection which assumes that individuals have a fundamental right to "be left alone".A plethora of studies across a variety of academic disciplines has examined the relationship between privacy concerns and privacy-related behaviours (Acquisti et al., 2015;Baruh and Popescu, 2017;Martin and Murphy, 2019).In the e-commerce context, it has been verified that perceived privacy concerns negatively affect usage behaviour (Vilmakumar et al., 2021).Moreover, privacy concerns are a major hindrance to data sharing in an online environment and limit the potential for e-commerce to grow; Angst and Agarwal (2009) and Buchanan et al. (2007) highlight that.In addition, it is argued that when consumers have privacy concerns or perceive threats to their personal information, consumers will withhold personal information, refuse participation in the online environment or refuse to disclose their personal information.
Consumer perceptions of privacy are an important influence on his or her information disclosure when consumers interact with a website (Lowry et al., 2012;Shah et al., 2014).Doubts about consumers' privacy result in consumer hesitancy to transact in e-commerce while privacy assurances are positively affecting consumers' behavioural intentions towards a website (Lowry et al., 2012).It has been suggested that consumers have a more positive perception towards information disclosure if privacy risks are low (Sun et al., 2015).A meta-analysis on 166 studies between 1990 and 2016 on privacy concerns found that privacy concerns predict information sharing (Baruh and Popescu, 2017).However, researchers have yet not adequately explained how consumers' attitudes around privacy influences the types of information disclosed and the level of disclosure (Pomfret et al., 2020).Another study finds that people who believe technology to be a potential threat to their privacy also demonstrate higher concerns about the technology itself, while people who believe it to be useful, will care less about privacy concerns (Vilmakumar et al., 2021).These

Role of anthro pomorphism
findings are in line with PCT.Consumers who have higher privacy concerns will perceive greater risks associated with their personal information being compromised or misused, which will affect their intentions to disclose private information negatively.Consequently, it is proposed that: RP3. Privacy concerns negatively affects consumers' intentions to disclose information.
Apart from that, privacy has been found to be an important variable regarding the adoption of innovative technologies (Kumar et al., 2016;Gutierrez et al., 2019).Privacy concerns also have a negative impact on consumers' attitude and intention to use an innovation (Müller-Seitz et al., 2009;Mani and Chouk, 2017) due to perceived intrusiveness.It can be observed that privacy concerns are negative attitudes or beliefs that influence an individual's attitudes, for instance, towards a technology, and usually centre around the collection of personal data, data errors, unauthorised access of personal data and unauthorised secondary use of personal data (Dinev et al., 2016).Thus, it is proposed that: RP4. Privacy concerns negatively affect consumers' attitude towards the digital assistant.
3.2.3Personalisation.AI agents such as Siri have become rapidly embraced and integrated into daily life by consumers because they enable individuals to access timely and useful information (Canbek and Mutlu, 2016).These agents meet customer demand for contextually relevant and highly personalised content that is delivered in real time (Brill et al., 2019).It is argued that companies use AI to collect user details with the aim to improve the user experience and enhance lifetime value of customers (Shankar, 2018;Wilson and Daugherty, 2018).Due to the circumstance, that with AI, data about every individual consumer can be stored and analysed at unprecedented scale, marketers can now personalise their marketing mix for each individual consumer.With the potential offered by AI, the goal of organisations is shifting now towards personalising interactions, delighting its audience and understanding each individual customer's unique challenges (Pansari and Kumar, 2017).Data sharing is an opportunity for consumers to participate in an exchange of data for retail value and personalized services (Barth and de Jong, 2017) Consequently, it is suggested that consumers react positively to personalisation because of improved experience and perceptions of being better understood by a company.It is hypothesized that this understanding will lead to: RP5.Personalisation positively affects consumers' intentions to disclose information.Personalisation has also led to higher click-through rates of advertisements (Aguirre et al., 2015;Bleier and Eisenbeiss, 2015).Personalisation results in more positive attitudes and higher click through rates.However, it should be noted that these studies also found that too much personalisation will impede trust and raise privacy concerns, which again reflects and highlights the relevance of the PPP.Nonetheless, it is proposed that: RP6.Personalisation positively affects consumers' attitude towards the digital assistant.

Attitude towards AI-based digital agents.
As suggested by the theory of planned behaviour (Ajzen, 1991), attitudes influence behavioural intentions.Above discussion has highlighted, how it is expected that privacy concerns, personalisation and anthropomorphism affect consumers¨attitude towards a conversational agent.Combining those arguments with Ajzen's theory, it is expected that attitudes towards AI has a mediating effect on the relationship between antecedents and intentions to disclose information to the conversational agent.Thus, it is proposed: RP7. Attitude towards the digital assistant mediates the effect of privacy concerns, personalisation and anthropomorphism on consumers' intention to disclose information.

Discussion
The review of existing literature has shown that consumers use AI technology in a variety of forms and for a variety of different tasks in their daily life (Brill et al., 2019).However, the success of AI depends on the availability of vast amounts of consumer data to offer benefits such as personalised content or discounts (Zhu et al., 2017).The PPP and PCT offer a valuable theoretical foundation in order to assess how AI can influence consumer information disclosure.The proposed research model and seven RPs suggest positive effects of anthropomorphism and personalisation on consumer attitude towards an intelligent digital assistant and positive effects on consumer information disclosure.Despite a lack of consensus across academic literature regarding the effects of anthropomorphism, it is argued that anthropomorphising AI will have positive effects such as encouraging consumers to like it more (Kim et al., 2019).Anthropomorphising AI by assigning it a name or character can nudge consumers to connect with them on a more personal level.This implies that if marketing managers are able to anthropomorphise their AI technology, consumers will evaluate them more favourably (Epley, 2018;Kim et al., 2019;Novak and Hoffman, 2019;Schweitzer et al., 2019).Anthropomorphism promises to overcome consumer hesitancy towards technology and engage with AI-based digital assistants on a personal level, offering companies opportunities for data collection and data curation for more meaningful value exchange and personalised offerings.
The next two RPs addressed the topic of privacy concerns relating to the collection, storage and analysis of consumer data through AI.Based on existing literature, it is suggested that privacy concerns negatively affect consumers evaluations of digital assistants, and negatively influence consumer intentions towards information disclosure to the digital assistant (Kumar et al., 2016;Gutierrez et al., 2019).It becomes evident that the topic of privacy can form a critical barrier to the success of AI-based digital assistants due to the circumstance that they depend on the presence and provision of personal and behavioural data in order to assist the consumer with personalised information during an interaction.Thomaz et al. (2020) content that society will witness a shift in the nature of the internet as privacy concerns and awareness about the use of private data is becoming more prevalent, which will significantly impact marketing as it benefits of wide availability of customer data as source of information for marketing strategy.Because marketing AI applications such as conversational agents are based on the use of personal information (Rai, 2020), it becomes necessary to examining the relationship of internet privacy concerns on consumer attitudes and consumer behaviour in this context of AI.
In addition to this, this study has drawn upon PPP and PCT to consider how personalisation offers benefits such as personalised content or discounts (Zhu et al., 2017) can potentially outweigh the perceived risks around personal data and privacy concerns and positively affect attitudes and intentions to disclose information.It is argued that through the provision of real-time, personalised and contextually relevant curation of information, Role of anthro pomorphism consumers evaluate digital assistants more positively and provide their information (Brill et al., 2019;Canbek and Mutlu, 2016).
Finally, this research highlights several gaps in academic literature relating to the effects of AI and the potential of AI to encourage consumer information disclosure.A detailed presentation of avenues for future research will be presented in Table 2 in the following section.

Conclusions, limitations and future research
Research into the effects of AI in a consumer context is still developing and sparse (Davenport et al., 2020).This research has addressed this gap and considered consumer information disclosure in the context of AI.A conceptual model with seven RPs has been developed, which can serve as foundation for future research.The following section will summarise insights from this paper and present a table with avenues for future research.

Conclusions
This study advances our understanding of the potential that AI offers in encouraging consumers to share their private information, which is critical for further advancements in AI that can only continue if companies have access to large amounts of data and computational capabilities to analyse them (Mazurek and Malagocka, 2019).Re-considering the benefits offered through increased personalisation against privacy concerns for the purpose of information disclosure in the light of AI-based digital assistants is critical due to the increasing integration of AI into marketing strategy and peoples' lives.Research pertaining to the effects of AI and AI-based technology is still at early stages and focuses heavily on technology adoption.Understanding the implications of AI in marketing and its impact on consumer behaviour becomes increasingly important to examine by researchers due to the circumstance that AI will systematically and effectively change the way in which companies connect and interact with their customers (Davenport et al., 2020;Grewal et al., 2020;Yadav and Pavlou, 2020).This research has developed a conceptual model and seven RPs that promise to a research gap and contribute to a better understanding of the implications of AI in the consumer sphere.What effect does the introduction of legislation such as GDPR have on consumers' privacy evaluations?Are consumers aware of regulation to protect their privacy?The "dark" side of AI invading privacy How can AI encourage consumers to share their data against their intentions?Can AI elicit information consumers are only subconsciously aware of?Complimentary AI technology (e.g.Siri as digital assistant) vs purchased AI technology (e.g.Amazon Echo) How does consumer attitudes and use differ for complimentary AI technology vs purchased AI technology?Do consumers perceive AI as invading their privacy?SJME

Limitations and future research
As with any research, this study is also subject to limitations that must be acknowledged and taken into consideration.Firstly, this research is conceptual in nature and has developed a research model (Ismagilova et al., 2020;Hughes et al., 2020;Kumar et al., 2018;Mathivathanan et al., 2022;Mishra et al., 2017;Rana et al., 2011Rana et al., , 2012) ) and RPs on an extant review of academic literature relating to AI, privacy calculus and consumer information disclosure.While this paper has offered new insights into the potential effects of AI on consumer attitudes and consumer information disclosure, it is even more important to test the proposed research model and its RPs.Hence, future research should consider the propositions presented in this research and test the conceptual model empirically.In addition, Table 2 presents areas for future research and potential research questions for future studies to consider.
Above table outlines three areas for future research.Firstly, future research can further investigate the PPP for the context of AI.For instance, research could ask how far personalisation can go before it becomes too intrusive and whether there are individual consumer characteristics, such as age, gender or culture, that influence this evaluation.Secondly, the European Union introduced GDPR (General Data Protection Regulation) in 2018, which aims to limit the way in which personal information can be stored and processed (Haenlein and Kaplan, 2019).The main task of this regulation is to strengthen the rights of the individual, making the expression of informed consent to disclose information mandatory as well as giving the opportunity to withdrawing consent (Mazurek and Malagocka, 2019).However, it is not clear whether consumers are actively aware of such legislation, or whether it influences their privacy and disclosure behaviour.Moreover, the potential "dark side" of AI and AI integration into marketing strategy is a research area that remains less studied (Grewal et al., 2021;Rana et al., 2022), consequently, there are opportunities for future research to investigate how AI could undermine consumer privacy.Finally, future research could differentiate between different types of AI technology.Finally, the future researchers can also review the literature (Alalwan et al., 2017;Alryalat et al., 2017;Dwivedi et al., 2017aDwivedi et al., , 2017bDwivedi et al., , 2021;;Kizgin et al., 2020;Patil et al., 2017;Rana et al., 2014) on the role of AI on consumers' information disclosure.
Figure 1.Proposed research model