Hitting the Jackpot With a Chatbot? Exploring the Determinants for a
Positive Attitude and Behavioral Intent Towards Chatbots Within the
Customer Service Industry
A study on the effect of perceived usefulness (PU), perceived ease of use (PEOU), privacy concern, and anthropomorphistic tendency on consumers’ attitude and behavioral intentions
towards chatbots in the customer service industry and the moderating role of anthropomorphistic tendency.
Eva Roos Teeuwissen (11274921)
Master Thesis – Graduate School of Communication
Master Program Communication Science: Corporate Communication Supervisor: Dr. Pytrik Schafraad
16th of October 2020 Wordcount: 7608
Abstract
To keep up with the digital transformation, companies are increasingly adapting their business and organizational activities, processes, and competencies to improve the way they conduct business. One way for them to digitalize their business is by deploying a chatbot as a new form of customer service. Nevertheless, as consumers are used to email or phone for interaction with an organization, changing to this new way of communication via a chatbot can be quite difficult. Therefore, analyzing significant sources which determine the attitude of users, and consequently intention to use, towards customer service chatbots is imperative. Moreover, chatbot research on privacy concern and anthropomorphistic tendency is still needed, even though they are considered important factors to influence consumers’ attitude. To get a better understanding on the functioning of chatbots this study examined the relationship between privacy concern and attitude, and consequently intention to use. Moreover, perceived usefulness and perceived ease of use are also included. Besides, this study acknowledges anthropomorphistic tendency as a moderator between the direct relationship of privacy concern and attitude. One hundred customer service chatbot users completed the online survey of this study. Finally, results reveal that for a chatbot to be perceived as something positive, perceived usefulness and perceived ease of ease need to be high. Unexpectedly, privacy concern did not influence attitude, and this relationship was not moderated by anthropomorphistic tendency. Also, high levels of anthropomorphistic tendency lead to a more negative attitude. Furthermore, attitude positively influences customers’ intention to use a chatbot.
Keywords: Chatbot, Customer Service, Perceived usefulness, Perceived ease of use, Privacy Concern, Anthropomorphistic Tendency, Attitude, Intention to Use.
Introduction
A well-known Dutch customer service chatbot example is that of the website Bol.com's chatbot Billie. It receives almost half of the customer contacts, and he successfully handles 70% of the questions (Boogert, 2019). Billie has already existed since 2008 and is unique in its connection with Bol.com's backend (Möller, 2018). Because of this, Billie can retrieve consumers' personal order information and has a pretty decent understanding of what the consumer wants to know. Billie can tell consumers that their package will arrive between five and seven this evening instead of telling them to log in to their account and look at the track & trace information. However, Billie is continually improving and will connect consumers to a real customer service agent when it doesn't understand their request.
Bol.com is one of many companies that digitize and automate their customer service to be more efficient due to the increasing availability and sophistication of software technologies such as chatbots (Benedikt & Osborne, 2013; Swanner, 2016). Chatbots are appealing for the customer service users as they increase productivity and allow for personalized 24/7
availability, therefore improving customer satisfaction (Androutsopoulou, Karacapilidis, Loukis, & Charalabidis, 2019; Ramachandran, 2019). Users can instantly get answers to their questions at any time, without waiting on a human. A survey by Kayak (2017) also highlights the importance of chatbots for customer service, as the results indicate that 64% use a chatbot to ask questions to the customer contact center of a company. Also, next to creating optimal customer satisfaction, chatbots allow businesses to reduce costs by up to 30% (McGrath, 2018).
Chatbots are often conceptualized as a computer program or conversational interface which makes use of natural language processing (NLP) to interact with people (Shawar & Atwell, 2007). They distribute contextual information and operate as a human-like customer
service representative to address consumers with information that fits the users’ needs at a suitable moment (Cui, Huang, Wei, Tan, Duan, and Zhou, 2017). This happens through the users’ preferred channel. According to McGrath (2018) as far as 80% of everyday customer
service questions can be answered by the chatbot. However, for this to happen, an organization's customers, before anything else, should carry favorable judgments about chatbots. Only then will they have the intention of actually using it.
A chatbot’s functionalities and users' expectations of these functionalities consequently form people's motivation for their attitude towards it. In turn, if the company’s chatbot will be successful or not all depends upon these attitudes. So, when chatbots provide inadequate responses to users' requests, this could lead to a gap between the users' expectations and the systems' performance. Customers will then experience unsatisfactory encounters with chatbots, and in turn, their attitude and usage intention towards chatbots will be negative. Therefore, it is essential that, for chatbots to be taken up and used as intended by the companies, chatbot users should have a favorable judgment with regard to the customer service chatbot (Adam, Wessel, & Benlian, 2020). Moreover, a positive attitude is crucial for a wide adoption of novel interactive solutions and has been argued to be a significant success factor in an online environment (Carlson & O'Cass, 2010).
A considerable array of previous research has already shown that users are more positive towards chatbots and are more likely to use them when they feel like the chatbot is useful and easy to use (Brandtzaeg & Følstad, 2017; Zarouali et al., 2018). This refers to the variables perceived usefulness and perceived ease of use, which are included in the
technology acceptance model theory (Davis, 1989). Davis (1989) and other scholars (Dabholkar & Bagozzi, 2002; Vijayasarathy, 2004; Weijters, Rangarajan, Falk &
Schillewaert, 2007) have established these concepts as essential determinants of technology adoption and are shown in studies of technology acceptance and as positive predictors of
attitude and intention to use. Therefore, this relationship is considered relevant to investigate by researchers, and acknowledging their importance they will also be included in this
research.
Even though these two important variables need to be included in this research, they do not completely predict the attitude consumers might have towards chatbots. Therefore, exploring other predictor variables is required (Wang, Wang, Lin & Tang, 2003). To receive personalized, valuable recommendations, personal information needs to be shared with the chatbot to gain relevant guidance. However, this might raise a certain level of privacy
concern among consumers and therefore influences users’ privacy-related actions like attitude and behavior (Dinev & Hart, 2006; Xu, Dinev, Smith, & Hart, 2008). Indeed, when
consumers experience a concern, they could assume a negative or positive perceptions, resulting in a positive or negative attitude (Ischen, Araujo, Voorveld, van Noort & Smit, 2019). Hereof, when interacting with digital technologies, chatbot usage might also enhance users’ privacy concerns (Dinev & Hart, 2006), which in turn could damage its' experience
with the organization (Ashworth & Free, 2006; Hong & Thong, 2013). Whether attitude and intention to use is influenced by the privacy concern of a customer service chatbot user is still researched to a limited extent in current literature. In today’s digital age, privacy concern evolved into a significant ethical issue, and it is therefore valuable for organization to comprehend if privacy concerns influence what users might think of the chatbot (Akhter, 2012; Phelps, Nowak & Ferrell, 2000; Saraty & Robertson, 2003; Wright & Xie, 2019). Therefore, this study looks into the privacy concern users might experience and if these privacy concerns might influence how users think about the chatbot and if they will consequently use it or not. In addition, when more knowledge is available about this
consent to data-collection. Accordingly, this study examines if privacy concern might influence how consumers feel towards chatbot and if they intend to use it.
Besides this, it is still unclear to what extent the individual differences in consumers' tendency to anthropomorphize a chatbot have a role to play in a customer service chatbot context. Previous chatbot research usually examines the effects of an anthropomorphized chatbot on attitude or privacy concern (Ischen et al., 2019; Araujo, 2018; Rietz, Benke & Maedche, 2019; Seeger, Pfeiffer & Heinzl, 2017). However, in reality, the chatbot user's individual differences may play a role and influence consumers' attitude and intention to use (Letheren, Martin, & Seung Jin, 2017). Accordingly, developers and marketeers will find value in more extensive knowledge about the anthropomorphizing tendency because it provides insights into several positive results. Encompassing emotional responses which might be favorable. Users attitude and behavior towards chatbot interactions might thus be impacted and consequently the success of a chatbot (Aggarwal & McGill, 2012). Thus, it should be considered to maximize the benefits of the human-robot interaction.
Other variables are included in this study that have not previously been included in such a context before and which might have an influence on and determine the attitude towards chatbots. Therefore, this research’s view on the attitude towards and usage of customer service chatbots is more extensive than what has been done before. Perceived usefulness and ease of use are considered to be two important factors in chatbot research and will thus also be included to get a better understanding on the functioning of chatbots
(Zarouali et al., 2018; Castañeda, Rios, & Martinez, 2007). The research gap on consumers’ privacy concerns and anthropomorphistic tendency will be filled by including these variables to get a better understanding on the functioning of chatbots. Hence, this research question was defined:
Research Question: To what extent is a users' attitude towards the chatbot, and consequently the intention to use influenced by perceived ease of use (PEOU), perceived usefulness (PU), and privacy concern. And how does anthropomorphistic tendency moderate the direct relationship between privacy concern and attitude?
Theoretical framework
This part of the study will discuss attitude, intention to use, perceived usefulness, perceived ease of use, privacy concern and anthropomorphistic tendency. At the end, Figure 1 will demonstrate the relationship between all concepts.
Attitude
"An attitude can be defined as an evaluative judgment, either favorable or
unfavorable, that an individual possesses and directs towards some attitude object" (Elias, smith & Barney, 2012, p. 454). Accordingly, for this research, the definition of attitude that will be used is the favorable or unfavorable judgment towards a customer service chatbot. A positive attitude acts as a valuable asset for an organization that wants to implement new technology, as a positive attitude indicates the consumers' intention to use it (Davis, 1993). Previous research shows that consumers develop a bias towards the attitude objects which they perceive as positive, but against these objects when they have a more negative evaluation (Eagly & Chaiken, 1998). Meaning, positive attitudes lead to the usage of
technology. However, this also means that a technology won’t be used when consumers have a more negative attitude (Liker & Sindi, 1997). Moreover, Culpan (1995) has stated, "there is a significant relationship between end-users' attitudes and their degree of command over the use of an information system. End-users must react favorably to a system to ensure that it will
be used widely and effectively" (p. 168). Therefore, the central aim of this study is to gain insights into the functioning of a chatbot, by looking at which variables in the customer service chatbot domain determine the attitude of a consumer.
Research has already shown that, a certain level of anxiety is developed when
consumers attitude towards a technology is negative (Marquié, Thon, & Baracat, 1994). This level of anxiety might then also create the perception that it is not easy to use. In turn, this could relate to how well users would welcome a technology that is recently put into operation (Fakun, 2009). Intensifying possible anxieties' impact which people might experience, and hence its' negative attitude. Thus, organizations benefit from a more positive attitude.
Intention to use
Fishbein and Ajzen (1974) define intention to use as "the strength of one's intention to perform a specified behavior" (p.288). This means that intentions are stimulated by certain determinants that influence motivation which in turn sway an individual’s behavior. This indicates the effort people are willing to put in to perform a certain behavior (Weijters et al., 2007). Prior literature shows an entrenched relationship between attitude and behavioral intention (Lee & Chang, 2011; Antón, Camarero & Rodríguez, 2013; Park, Baek, Ohm, & Chang, 2014; Herbjørn, Pedersen & Thorbjørnsen, 2005). This relationship entails that when technology users have a positive attitude, they are more likely to use the technology (Zarouali et al., 2018). This assumption agrees with other research which demonstrates the link
between users’ technology usage intent and their actual behavior to actually use it
(Bhattacherjee & Sanford, 2009). Taking this in mind, the results of Rice and Aydin (1991) were to be expected, in which they established that in order to research the implementation of technology there should be a fundamental understanding of the attitude towards this
Consequently, this research pursues the same direction and expect attitude to be positively related to consumers’ intention to use customer service chatbots. Hence,
hypothesis 6 was defined:
Hypothesis 6 (H6): There is a positive direct relationship between attitude and intention to
use.
Perceived usefulness
Perceived usefulness (PU) is defined as "the degree to which a user believes a
particular technology enhances performance" (Davis et al., 1989, p.985). Specifically, when a chatbot’s performance is quick and effective. Consumers base their acceptance of a certain
technology on the benefits they receive when using the technology, are the benefits high they will accept the technology and vice versa. Perceived usefulness is therefore shown to be the strongest motivation for technology acceptance is various research (Kulviwat, Songpol, Bruner, Gordon, Kumar, Anand, Altobello, Terry and Terry, 2007). Additionally, consumers' brand experience and attitude towards new digital technologies have been positively
influenced (Aaker & Jacobson, 2001; Davis et al., 1989; Lee & Chang, 2011).
As attitude toward a brand is a significant predictor on consumers intent to use it, it is important brand attitude is positive (Zarouali et al., 2018). This is further supported by the research of Abdul (2019) in which attitude is positively influenced by perceived usefulness and behavioral intent is positive affected by a positive attitude. Moreover, the study by Brandtzaeg & Følstad (2017) suggests that productivity is the most frequently suggested motivational factor, since users feel that chatbots help them get timely and efficient assistance. This can be translated into perceived usefulness.
In previous research perceived usefulness is found to have a positive influence on customers’ attitude with regard to chatbots and other technologies (Dabholkar & Bagozzi,
2002; Vijayasarathy, 2004; Weijters et al., 2007; van Eeuwen, 2017), this study suggests the same positive relationship is true for chatbots. Hence, hypothesis 1 was defined:
Hypothesis 1 (H1): There is a positive direct relationship between perceived usefulness
(PU) and users' attitude towards the chatbot.
Perceived ease of use
Perceived ease of use (PEOU) is defined as "the degree to which the prospective user
expects that target system to be free of effort" (Davis et al., 1989, p.985). Meaning that consumer will not experience any effort when using a chatbot (Zarouali et al., 2018). Adding to that, PEOU is a component of the so-called cost-benefit trade-off, where the costs are seen as the amount of effort for the consumer to put in (Johnson & Payne, 1985; Maslowsky, Buvinger, Keating, Steinberg, & Cauffman, 2011). A low effort expectation arises when chatbot users have a high ease of use perception (Ellis & Kurniawan, 2000). Ease of use has a direct association with a users’ attitude as efficiency is seen as something that is personally rewarding (Castañeda, Rios, & Martinez, 2007). Previous research has acknowledged PEOU as an important variable determining a consumers’ attitude. This attitude can go from buying clothes online, (Vijayasarathy, 2004; Schierz, Schilke & Wirtz, 2010), and also consumers' general attitude toward SNS sites (Lorenzo, Constantinides, & Alarcon-del-Amo, 2011). However, more importantly, Zarouali et al. (2018) explicitly indicate that when consumers experience a positive PEOU, their attitude towards the chatbot brand will also be more positive. Thus, a high score on PEOU indicates the chatbot to be perceived as easy to use renders into the expectation of consumers perceiving the chatbot as easy to use, accordingly adopt chatbots quicker than consumers who perceive chatbots as difficult to use (Zarouali et al., 2018). Therefore, this research expects that PEOU will positively predict attitude towards the chatbot. Hence, hypothesis 2 was defined:
Hypothesis 2 (H2): There is a positive direct relationship between perceived ease of use
(PEOU) and users' attitude towards the chatbot.
Privacy concern
Privacy concern is a very broad concept. Literature indicates privacy concern to be a concept with different understandings and in different contexts. According to Taylor,
Ferguson, and Ellen (2015), privacy can be seen as either a good that has economic value or a human right that must be protected, which they describe as a value-based definition. They also describe a cognate-based definition, when someone behaves in a certain way like when they have the feeling that they are in control (Taylor et al., 2015).
Bansal, Zahedi, and Gefen (2016) suggest that when consumers do business in an online environment trust is a valuable aspect for them. Moreover, risk is also important, which is the possibility that a negative outcome will occur. This is important because negative outcomes can influence an individual emotionally, materially, and physically (Norberg, Horne & Horne, 2007).
Privacy concern, in turn, is related to information sharing and the unfavorable results it can bring (Baruh, Secinti, & Cemalcilar, 2017). When consumers experience privacy concern it can be related to the obtained information about them as a person, the feeling of losing control over their information, and the understanding of how privacy is exercised. When consumers are burdened with these worries, they will make more calculated decisions about their data sharing (Zukowski & Brown, 2007). This research will follow the definition used by Ischen et al. (2019); "The degree to which a consumer is worried about the potential invasion of the right to prevent the disclosure of personal information to others" (p. 3). Consumers' disclosure choice depends upon the privacy concerns of the information owner (Shibchurn & Yan, 2015). Information is more readily disclosed when the consumers'
impression is that the offered benefits from disclosing outweigh the risks (Shibchurn & Yan, 2015). For customer service chatbots to make recommendations, they need to collect personal information. These can be collected through direct requests for information disclosure, which in turn might induce privacy concern (Wottricht, Verlegh & Smit, 2017). Previous research by Følstad, Nordheim & Bjørkli (2018) indicated that when consumers interact with chatbots, they have a concern for privacy. They argue that for some consumers chatbots might still be a new phenomenon, meaning that they could be more aware of the requests for information, which in turn could induce their privacy concern. Furthermore, an important antecedent for mobile chatbot acceptance is privacy concern (van Eeuwen, 2017).
Results of previous research by Baruh and Cemalcilar (2017) show that individuals’ privacy concern will determine how they use online services and deal with how they manage their privacy (i.e., information sharing). These results are in line with the assumptions of the privacy paradox. Meaning that distinctions between privacy concerns and behavior may relate to user motivations and the associated risk-benefit analysis. Users perception of privacy risks might be lower than their perceived benefits, when sharing information (Baruh &
Cemalcilar, 2017).
However, more research shows the opposite relation in which more privacy concern leads to a more negative attitude. One relevant study on the antecedents of mobile chatbot acceptance is that of van Eeuwen (2017). Results of this study show that mobile chatbots are more likely to be accepted when privacy concerns are low, therefore having an influence on the attitude of the chatbot user. This negative relationship exists because consumers are concerned about if the company behind the chatbot might use their personal information for economic self-interest (van Eeuwen, 2017). Results by Dinev and Hart (2006) suggest that same relationship in which high privacy concern results in consumers being less willing to share personal information when having to make transactions. An explanation for this
relationship was given by Dinev and Hart (2006): consumers are less certain about which individuals can see their personal information and in what way they are using it, therefore their privacy concerns are higher and they are less willing to share information. Consumers experience this because of the uncertainties the technology itself brings. Consumers
experience a more negative attitude and a lower intention to purchase if their privacy concerns are high (Kim & Huh, 2017; Phelps, D'Souza & Nowak, 2000). As purchase intention is a behavioral intention, this research also supports the negative relationship between privacy concern, consumers' attitude, and behavioral intent. Lastly, when consumers are more worried about sharing personal information they can become skeptic, which in turn might result in a more negative attitude (Kim & Huh, 2017). So, when chatbots request personal information, consumers' privacy concern might be induced (Wottrich et al., 2017). Consistent with research by Følstad et al. (2018), whose results present a negative attitude when they have a high privacy concern when interacting with chatbots. Following this reasoning, the following hypothesis was formulated:
Hypothesis 3 (H3): There is a negative direct relationship between privacy concern and
users' attitude towards the chatbot.
Anthropomorphistic tendency
Anthropomorphistic tendency is defined as the "individual propensity to utilize the process of anthropomorphism" (Letheren, Martin & Jin, 2017, p.67). This concept suggests that anthropomorphism varies by individual. Therefore, chatbot users are more inclined to perceive a chatbot as humanlike when their anthropomorphistic tendency is high, thereby utilizing anthropomorphism (Waytz et al., 2010). Moreover, consumers with a high
anthropomorphistic tendency are expected to have a more positive attitude. This comes from the notion that when users consider something to be human, they consider it to be very much
alike. This is important because people appreciate other people/things which are alike. Meaning that feeling a chatbot is humanlike makes it look more like the human perceiver (Izard, 1960; Byrne, Griffitt & Stefaniak, 1967). Aggarwal and McGill (2007) further support this as their results show that objects are seen as more favorable when people possess a high anthropomorphistic tendency.
The chatbot and the chatbot user, experience three repercussions when an
anthropomorphistic tendency occurs (Epley & Waytz, 2009). The responsibility, trust, and social influence a chatbot receives depends upon the individual differences created by the anthropomorphistic tendency.
According to Go and Sundar (2019) the social connectedness induced by
anthropomorphized chatbot allows a consumer to be more positive about the chatbot and also more likely to use it. So, when people have a high tendency to anthropomorphize, they feel like they are interacting with a humanlike chatbot, which can positively influence attitude. Moreover, when people perceive a chatbot to possess a consciousness, a chatbot is able to perform a deliberate action (Gray, Gray & Wegner, 2007). Therefore, we assign responsibility to the chatbot. If a person can decide for oneself that an agent can be blamed, then likewise, it is also possible for an agent to be perceived as trustworthy. Therefore, significant information can be trusted upon technology by those whose anthropomorphistic tendency is high. Which could be trusting the chatbot to handle personal information, therefore having a lower privacy concern. Another experiment also illustrates the previous findings (Hinds, Roberts & Jones, 2004): Some people of the experiment were instructed to give anthropomorphistic characteristics to a robot with which they had to work. More
responsibility was assigned to the robot who did this work than to the robot of the group who weren’t instructed to anthropomorphize it (Hinds, Roberts & Jones, 2004; Burgoon et al.,
The moderating relationship of anthropomorphistic tendency is further supported in the study of Spiekermann, Grossklags & Berendt (2001). They take another perspective on the usage of chatbots and its' relation to privacy concern. Their online shopping experiment, in which they enter a dialogue with an 'anthropomorphic 3-D shopping bot' (chatbot), nicely pictures an example. This research contrasted most of the other research on privacy concern and chatbot usage. The research showed a high privacy concern among almost all participants but readily revealed a large amount of information when talking with the chatbot. Even a differentiation in the privacy statement that was included did not impact how much
information the participants shared with the chatbot (Spiekermann, Grossklags & Berendt, 2001). These results indicated evidence for the privacy paradox in their research of
information disclosure to a chatbot in relation to their privacy concerns. When participants were asked why they revealed this information to the chatbot, they said they 'felt personally addressed' by the chatbot and liked her type of 'soft communication.' This might be an explanation of why users disclose information differently to anthropomorphized chatbots. Therefore, we expect that the tendency to anthropomorphize will act as a moderator for the relationship between privacy concern and attitude. Moreover, a high tendency to anthropomorphize will result in a more positive attitude. The last hypotheses were thus formulated as follows:
Hypothesis 4 (H4): Anthropomorphistic tendency moderates the relationship between
privacy concern and users' attitude towards the chatbot.
Hypothesis 5 (H5): There is a positive direct relationship between anthropomorphistic
Figure 1. Conceptual Model
Method
Research design
A cross-sectional online survey was used to collect data to examine the relationship between PEOU, PU, privacy concern, anthropomorphistic tendency and attitude and intention to use the chatbots.. There are two reasons for the use of a cross-sectional survey: it allowed us to study a large population at one point in time and also allowed us to draw generalizable conclusions. Moreover, this modality comes with a low cost, serving as another benefit. A self-report measure was used to gain responses for the survey in which each participant
independently filled in the questions. The survey began with an informed consent that the respondents had to accept before continuing the survey.
Data collection and sample
The personal network of the researcher was used to distribute the online survey. The following channels were used: e-mail, LinkedIn, Facebook, and Whatsapp. The participants in the sample should be 18 years or older and have used a chatbot before to be able to fill in the survey. Appendix I includes the full survey.
This research used a non-probability convenience sampling method since the participants of this study were all conveniently available to the researcher. The reason for this sampling method was the collection of a bigger sample.
A total of 120 people began the questionnaire and 20 of them indicated that they have never used a chatbot before. This means 20 responses were deleted because they have never used a chatbot before which means they are not valid for this research. Meaning that the total sample consisted of 100 participants. In total, 57 females (57%) and 43 males (43%) participants make up the total sample. The age of the participants ranged from 18 to 56 years (M=29, SD=11). Furthermore, 21% of the participants have obtained a secondary degree, 55% a bachelor’s degree and 24% a master’s degree. Finally, 93% of the participants said that they use the chatbot less than once per month.
Variables and measures
All the variables and measures relate to customer service chatbots. Before starting the survey, respondents had to indicate what customer service chatbot they have used before and were told to relate all the questions back to this, with which they have an experience.
Attitude
As stated before, attitude is conceptualized as “an evaluative judgment, either favourable or unfavourable, that an individual possesses and directs towards some attitude object” (Zaphiris & Ioannou, 2018, p.27). The measure of attitude consisted of only four items and was used to measure the first dependent variable. A seven-point Likert scale was used to measure attitude ranging from (1) strongly disagree to (7) strongly agree (Venkatesh, Morris, Davis, and Davis, 2003). An example of such a statement is: “using chatbots is a good idea”. A principal axis factoring analysis was conducted for engagement with oblique rotation (direct Oblimin) to test the correlation. The sampling adequacy was middling, KMO = .70. One factor had an eigenvalue over Kaiser’s criterion of 1 and explained 79.33% of the variance (Appendix C, table 5). The 4-item scale also proved excellent reliability with a Cronbach’s Alpha (α) of .91. The total score of ‘attitude’ was computed by using the mean across the four items (M = 3.25, SD = 1.45).
Intention to use
As stated before, intention to use is conceptualized as “an evaluative judgment, either favourable or unfavourable, that an individual possesses and directs towards some attitude object” (Zaphiris & Ioannou, 2018, p.27). The measure of intention to use consisted of only four items and was used to measure the second dependent variable. The four items were measured on a seven-point Likert scale ranging from (1) strongly disagree to (7) strongly agree (Venkatesh, Morris, Davis & Davis, 2003). An example of such a statement is: “using chatbots is a good idea”.
A principal axis factoring analysis was conducted for engagement with oblique rotation (direct Oblimin) to test the correlation. The sampling adequacy was middling, KMO = .73. One
factor had an eigenvalue over Kaiser’s criterion of 1 and explained 76.59% of the variance (Appendix C, table 6). The 4-item scale also proved good reliability with a Cronbach’s Alpha (α) of .89. The total score of ‘intention to use’ was computed by using the mean across the four items (M = 3.33, SD = 1.53).
Perceived usefulness
As stated earlier, the perceived usefulness is conceptualized as the perceived likelihood that the chatbot will enhance a consumer’s productivity or job performance (Davis et al., 1989). A perceived usefulness measure was used to measure this independent variable. The measure consisted of five items which were all measured on a seven-point Likert scale. The scale ranged from (1) strongly disagree to (7) strongly agree (Davis et al., 1989). An example of such a statement is: “communicating with a chatbot saves me time”. All five statement can be found in Appendix B.
A principal axis factoring analysis was conducted for engagement with oblique rotation (direct Oblimin) to test the correlation. The sampling adequacy was mediocre, KMO = .65. One factor had an eigenvalue over Kaiser’s criterion of 1 and explained 63,04% of the variance (Appendix C, table 1). The 4-item scale also proved good reliability with a Cronbach’s Alpha (α) of .80. The total score of ‘perceived usefulness’ was computed by using the mean across the four items (M = 3.32, SD = 1.15).
Perceived ease of use
As stated earlier, the perceived ease of use is conceptualized as the degree to which consumers think that using the chatbot will be simple and free of any effort (Zarouali et al., 2018). The perceived ease of use measure consisted of seven items and was used to measure the independent variable perceived ease of use. The measure used a seven-point Likert scale
ranging from (1) strongly disagree to (7) strongly agree (Davis et al., 1989). An example of such a statement is: “overall, I find chatbots ease to use”. Appendix B includes all seven statements.
First, item 1,2,3,5, and 6 needed to be reversed coded for this variable. After, a principal axis factoring analysis was conducted for engagement with oblique rotation (direct Oblimin) to test the correlation. The sampling adequacy was meritorious, KMO = .84. One factor had an eigenvalue over Kaiser’s criterion of 1 and explained 63,04% of the variance (Appendix C, table 2). The 4-item scale also proved excellent reliability with a Cronbach’s Alpha (α) of .90. The total score of ‘perceived ease of use’ was computed by using the mean across the four items (M = 2.87, SD = 1.24).
Privacy concern
As stated earlier, the privacy concern is conceptualized as “the degree to which a consumer is worried about the potential invasion of the right to prevent the disclosure of personal information to others” (Baek & Morimoto, 2012 p.63). The privacy concern measure of Xu et al. (2008) was used to measure this independent variable. The measure consisted of five items, which were measured on a seven-point Likert scale which ranges from (1) strongly disagree to (7) strongly agree (Xu et al., 2008). An example of such a statement is: “It bothers me when chatbots ask me for personal information”. Appendix B includes all statements.
A principal axis factoring analysis was conducted for engagement with oblique rotation (direct Oblimin) to test the correlation. The sampling adequacy was meritorious, KMO = .84. One factor had an eigenvalue over Kaiser’s criterion of 1 and explained 87,44% of the variance (Appendix C, table 3). The 4-item scale also proved excellent reliability with a Cronbach’s Alpha (α) of .96. The total score of ‘privacy concern’ was computed by using the mean across the four items (M = 3.49, SD = 1.83).
Anthropomorphic tendency
The variable anthropomorphistic tendency is a moderating continuous variable in this research. This variable indicates the individual propensity towards anthropomorphism (Waytz, Cacioppo & Epley, 2010). The IDAQ scale from Waytz, Cacioppo and Epley (2007) was used in order to measure the level of anthropomorphistic tendency, five items were chosen to measure anthropomorphism of technology. A ten-point Likert scale was used in order to measure the five items ranging from (1) not at all to (10) very much. An example of a statement measuring anthropomorphistic tendency is: “to what extent does the average robot have consciousness?”. Appendix B includes all five items.
A principal axis factoring analysis was conducted for engagement with oblique rotation (direct Oblimin) to test the correlation. The sampling adequacy was middling, KMO = .73. One factor had an eigenvalue over Kaiser’s criterion of 1 and explained 76.85% of the variance (Appendix C, table 4). The 4-item scale also proved excellent reliability with a Cronbach’s Alpha (α) of .91. The total score of ‘privacy concern’ was computed by using the mean across the four items (M = 4.40, SD = 2.36).
Demographic variables and covariates
This study included three demographic variables: gender, age, and education. Gender consisted of male, female, and other. The education level options were none, elementary school, secondary education, bachelor’s degree, master’s degree, doctorate degree, and other. Besides these demographic variables, a five-point scale was used in order to measure the participants’ chatbot use. This ranged from (1) = never and (5) = daily. The average level of anthropomorphistic tendency of the respondents was quite low ( M = 4.4).
An overview of all the customer service chatbots that have been used by the participants can be found below. Some notable customer service chatbots, that have been used more often by respondents were the chatbot from Bol.com (11%), Coolblue (5%), ING (8%), and Rabobank (11%). The full overview of all the customer service chatbots can be found in Appendix C, table 10.
Results
Multiple regression analysis was computed to test hypotheses H1, H2, H3, and H5. The reason for choosing this method is twofold. First of all, through a multiple regression the relative relationship between the attributes can be examined to see which attribute has the strongest effect. Moreover, the influence of one attribute can be examined when the others are kept constant.
The model as a whole is significant, F(4,95) = 58.19, p < .001. Perceived usefulness (b*= .52, t = 6.93, p < .001), and perceived ease of use (b*= .36, t =4.76, p <.001) significantly predict attitude, but the effect of privacy concern is not significant (b*= -.04, t = -.76, p = .45). These variables together explain 71% of the variance in attitude (R2 = .71). In this model, perceived usefulness has the strongest effect on attitude (Appendix C, table 7). So, hypothesis H1 and H1 are supported in this analysis and H3 is rejected. Anthropomorphistic tendency significantly negatively predicts attitude (b*= -.14, t =-2.39, p = .02). So, H5 needs to be rejected, since a positive relation was expected. These results can be found in Appendix C, table 7.
In order to test the fourth hypothesis “Anthropomorphistic tendency moderates the relationship between privacy concern and users’ attitude towards the chatbot”, we tested the interaction of anthropomorphistic tendency within a multiple regression analysis via
PROCESS model 1. So, we want to address whether the relationship between privacy concern and attitude differs for people with a different level of anthropomorphistic tendency.
The results show that the regression model was significant, F (1,100) = 2,94, p < .05. the regression model could therefore be used to predict the attitude, the strength of the prediction is very weak: 8,4% of the variation in attitude could be predicted on the basis of privacy concern and the level of anthropomorphistic tendency (R2 = .084). As already shown in the multiple regression, privacy concern did not significantly affect the attitude. Privacy concern, b* = -0.07, t = -0.68, p = .496, 95% CI [-0.17, 0.08], has no significant association with attitude. It did appear that anthropomorphistic tendency had a significant strong association with attitude, b* = -0.28, t = -2.92, p < .05, 95% CI [-0.23, -0.04]. However, it appeared that the interaction effect was not significant, and anthropomorphistic tendency has no statistically significant positive interaction effect on a person’s attitude at average privacy concern b*= 0.002, t = 0.32, p = .752, 95% CI [-0.01; 0.01]. Thus, the fourth hypothesis must be rejected. Thus, the effect of privacy concern on a persons’ attitude is the same for all levels of a persons’ anthropomorphistic tendency. There is no moderating effect of anthropomorphistic tendency in the relationship between privacy concern and attitude. The results can be found in Appendix C, table 9.
Finally, in order to test the sixth and last hypothesis “There is a positive direct relationship between users’ attitude towards chatbots and users’ intention to use chatbots” a simple regression analysis is computed to test the relationship between attitude and intention to use. The regression model for attitude was significant F (1,98) = 167.73, p < .001. Attitude has a positive significant association with intention to use, b* = 0.79, t = 12,95, p < .001, 95% CI [0.71, 0.97] and explains 63,1% of the variance in transparency (R2 = .63). For every higher level of attitude, the level of intention to use increases with .79. Thus, this analysis supports H7; attitude predicts intention to use. The results can be found in appendix C, table 8.
Discussion
This research aimed to provide more understanding into the functioning of chatbots, through examining the research question: To what extent is a users' attitude towards the chatbot, and consequently the intention to use influenced by perceived ease of use (PEOU), perceived usefulness (PU), and privacy concern. And how does anthropomorphistic tendency moderate the direct relationship between privacy concern and attitude? The literature about
TAM (David, 1989) is supported through the incorporation of the variables PU and PEOU. Moreover, the literature is extended through the inclusion of the variables privacy concern, anthropomorphistic tendency and their relation to attitude. Also, including
anthropomorphistic tendency as a moderator of the relationship between privacy concern and attitude.
Previous research is supported through the following result: there is a positive direct relationship between PU, PEOU and attitude (Dabholkar & Bagozzi, 2002; Vijayasarathy, 2004; Weijters et al., 2007; van Eeuwen, 2017). These results confirm that functionality is an important factor to consider when implementing a chatbot. Therefore, organizations need to make sure their chatbot enhances a consumers' performance and make the chatbot free of any effort. They have to let the chatbot respond to the consumer quickly and in a time-efficient manner to increase usefulness. Moreover, the variable that predicted attitude the most was PU. This means that consumers will have a more positive attitude when the chatbot enhances their performance, rather than feeling the chatbot to be free of any effort. This supports the idea that perceived usefulness strongly correlates with customers' acceptance of the
technology (Davis, 1989). This relationship could be explained because of the importance consumers' attribute to the benefit from an innovation or if they won't (Kulviwat et al., 2007) and the way consumers give value to the functions of the chatbot when they are considering
adopting this technology (Davis, 1989). This may especially be critical for the customer service domain, as consumers want to use the chatbot to quickly gain information. In contradiction to the expectation of this study, privacy concern did not predict attitude, and H3 had to be refuted. So, it was not the case that attitude was more negative when privacy concern was higher. This result is inconsistent with the results of previous research, these results suggest that consumers attitude and intent is negatively influenced by privacy concern by showing higher skepticism and therefore have a more negative attitude (Kim & Huh, 2017). The first explanation for this could be that the benefits of sharing information might outweigh the concerns consumers have and thus still have quite a positive attitude. Moreover, another possible explanation for this result could be due to two of the assumptions which demonstrate unfavorable results on privacy concern which could occur when information is shared. These unfavorable results which occur when information is shared is the feeling of how much control consumers experience, and how conscious they are of the privacy practices. Within customer service chatbot use, consumers are mostly aware of sharing the information. Therefore, privacy concern might not be as bad in this context. Also, because they are aware, their potential loss of control might be mitigated. Therefore, for future research, another dimension of privacy concern could be considered to see if this better predicts attitude. There are many other dimensions within privacy concern which could be considered for example, if transactions are supported by chatbot, instead of only FAQ’s (Følstad, Nordheim & Bjørkli, 2018).
Again, as a contradiction to the expectations of this study, the direct relationship between privacy concern and attitude was not moderated by the difference in
anthropomorphistic tendency, so H5 needed to be rejected. It was assumed that if a consumer had a high anthropomorphistic tendency, the direct relationship between privacy concern and
attitude would be weaker. This result occurs because this direct relationship was not significant to begin with.
Hypothesis five shows interesting findings. It appeared that even though the
relationship between privacy concern and attitude was not moderated by anthropomorphistic tendency, there was a direct effect of anthropomorphistic tendency on attitude. However, unexpectedly, this study showed that anthropomorphistic tendency negatively influenced the attitude. So, it appeared that people with a higher tendency to anthropomorphize had a more negative attitude. That is why the fifth hypothesis was rejected. This could be explained by the fact that when consumers have a high anthropomorphistic tendency they believe that all the chatbot’s actions are deliberate. They could feel like the chatbot could intentionally
misuse information, and this would mean that consumers perceive the risk to privacy as higher than the benefits of sharing information. The conclusion will include organizational implications. Furthermore, different individual factors might have an influence on
anthropomorphistic tendency and should therefore be taken into future research. This way organizations could gain insight into what kind of consumers they could target in what way. For example, if age or culture might affect individual anthropomorphism and if this in turn influences attitude. It was predicted that, on the basis of past research, intention to use would be positively influenced by attitude (Davis, 1989). Confirming the final hypothesis.
Finally, the overall research question can be answered. A more positive attitude towards chatbots is achieved through a high PU and PEOU. Therefore, organizations need to create a chatbot which provides productive services and are free of any effort because this will create a positive chatbot attitude and consumers will actually use it. Moreover,
organizations do not need to worry about the privacy concerns consumers' might experience as they will not have any influence on consumers' attitudes. Furthermore, this relationship is not necessarily different for consumers with different levels of anthropomorphistic tendency.
A few limitations are noteworthy when observing this study. To begin, the study made use of a small sample size (N = 100). Also, the general population is not fully represented in the sample, making the results not generalizable. A majority of the sample has a WO
bachelor or master, meaning that the majority of the sample has a high education. By having more variation in the sample in future research the external validity is improved. Moreover, for future research the survey should also be translated into Dutch because then a larger sample can be reached, compared to a survey that is only in English. Hereby people who only speak Dutch will also be included in the survey.
What is more, the single domain this research focuses on is the customer service domain. Meaning that the results are only generalizable for the customer service domain and not to others. As stated before, it is assumed that chatbots enhance customer support and subsequently, customer satisfaction (Ramachandran, 2019), and that is why the focus in this study was on the customer service domain. Nevertheless, it would also be important for chatbot research to look at other domains and if there is any influence on attitude in the future, such as onboarding or sales support. This might also explain the results concerning privacy concern, stating that privacy concern does not affect attitude. This might be different, considering different domains.
Another limitation is that 93% of participants indicated their chatbot use to be less than once per month. The limited chatbot use might influence the results of this study, where people who use chatbots more often might have a better-developed opinion and thus provide different insights. Future research should further examine if the amount of chatbot use might affect variables such as privacy concern, attitude, and intention to use.
Finally, this research only affords an understanding of the consumer’s future intention to use the chatbot. This does, however, not show the actual behavior of customers. A
Conclusion
In conclusion, by looking at the relationship between PU, PEOU, and attitude this research extends the current literature. Also, including the direct relationship between attitude and intention to use. Moreover, the research’s importance lays in its contributes to the
literature by including the relationship between privacy concern, anthropomorphistic tendency and attitude.
Moreover, this research also has managerial implications. An implication that is once again proved by this research is that the implementation of a chatbot for an organization should consider its’ perceived usefulness and perceived ease of use. Most importantly,
however, they should focus on the perceived usefulness, as functionality is most important. Furthermore, organizations do not always need to consider consumers’ privacy concerns as they do not necessarily influence their attitude and behavior towards the chatbot. Meaning that companies and policymakers might not need to include any privacy disclosure of their information/privacy setting before chatting. As consumers with a higher tendency to
anthropomorphize are shown to have a more negative attitude, organizations could consider showing consumers what the chatbot will do with their collected information so that
consumers could be reassured that they will not intentionally misuse information.
This research provides insights on, the attitudes and behavioral intentions, consumers have on chatbots use in customer service. Striving for a positive attitude and use of chatbots is not solely based upon the PU and PEOU of the chatbot. In addition to that consumers’ tendency toward attributing human-like characteristics to the chatbot. Privacy concern does not predict attitude but could certainly still be an important indicator of attitude.
Aaker, D. A., & Jacobson, R. (2001). The value relevance of brand attitude in high-technology markets. Journal of Marketing Research, 485-493.
Adam, M., Wessel, M. and Benlian, A. (2020), “AI-based chatbots in customer service and their effects on user compliance”, Electronic Markets, Vol. 9 No. 2, p. 204
Akhter, S. H. (2012). Who spends more online? The influence of time, usage variety, and privacy concern on online spending. Journal of Retailing and Consumer Services, 109-115.
Androutsopoulou, A., Karacapilidis, N., Loukis, E., & Charalabidis, Y. (2019). Transforming the communication between citizens and government through AI-guided chatbots. Government Information Quarterly, 358-367.
Antón, C., Camarero, C., & Rodríguez, J. (2013). Usefulness, enjoyment, and self-image congruence: the adoption of e-book readers. Psychology & Marketing, 372-384. Araujo, T. (2018). Living up to the chatbot hype: The influence of anthropomorphic design
cues and communicative agency framing on conversational agent and company perceptions. Computers in Human Behavior, 183-189.
Ashworth, L., Free, C. (2006). Marketing dataveillance and digital privacy: using theories of justice to understand consumers’ online privacy concerns. J Bus Ethics, 107–123 https://doi.org/10.1007/s10551-006-9007-7
Bansal, G., Zahedi, F., Gefen, D. (2015). Do context and personality matter? Trust and privacy concerns in disclosing private information online. Information &
Benedikt, C., Osborne, M. (2013). The Future of Employment. Retrieved from Oxford Martin Programme on Technology and Employment:
http://sep4u.gr/wp-content/uploads/The_Future_of_Employment_ox_2013.pdf
Bhattacherjee, A., & Sanford, C. (2009). The intention-behaviour gap in technology usage: The moderating role of attitude strength. Behaviour & Information Technology, 28(4), 389–401. https://doi.org/10.1080/01449290802121230
Boogert, E. (2019, January 25). Billie van bol.com doet 45% klantcontact. Retrieved from Emerce: https://www.emerce.nl/nieuws/billie-van-bol-com-doet-45-klantcontact Brandtzaeg P.B., Følstad A. (2017) Why people use chatbots. In: Kompatsiaris I. et al.
(eds) Internet Science. INSCI 2017. Lecture Notes in Computer Science, vol 10673. Springer, Cham
Burgoon, J., Bonito, J., Bengtsson, B., Cederberg, C., Lundeberg, M., Allspach, L. (2000) Interactivity in human-computer interaction: A study of credibility, understanding, and influence. Computers in Human Behavior. 16. 553-574. 10.1016/S0747-5632(00)00029-7.
Byrne D, Griffitt W, Stefaniak D. Attraction and similarity of personality characteristics. J Pers Soc Psychol. 1967;5(1):82-90. doi:10.1037/h0021198
Castañeda, A., Ríos, F.J.M., and T.L., Martínez (2007). The dimensionality of customer privacy concern on the internet. Online Information Review. 31. 420-439. 10.1108/14684520710780395.
Carlson, J., & O'Cass, A. (2010). Exploring the relationships between e-service quality, satisfaction, attitudes and behaviours in content-driven e-service web sites. Journal of Services Marketing, 112-127.
Computational Linguistics-System Demonstrations (pp. 97-102). Vancouver: Association for Computational Linguistics.
Culpan, O. (1995). Attitudes of end-users towards information technology in manufacturing and service industries . Information & Management , 167-176.
Dabholkar, P. A., & Bagozzi, R. P. (2002). An attitudinal model of technology-based self-service: moderating effects of consumer traits and situational factors. Journal of the Academy of Marketing Science, 184-201.
Davis, F. D. (1989). Perceived usefulness, perceived ease of use, and user acceptance of information technology. MIS Quarterly, 319-340. doi: 10.2307/249008
Dinev, T., & Hart, P. (2006). An extended privacy calculus model for e-commerce transactions. Information Systems Research, 61-80.
Eagly, A. H., & Chaiken, S. (1998). Attitude structure and function. In D. T. Gilbert, S. T. Fiske, & G. Lindzey (Eds.), The handbook of social psychology (p. 269–322). Boston: McGraw-Hill.
Eeuwen, M. van (2017) Mobile conversational commerce: messenger chatbots as the next interface between businesses and consumers. Retrieved from:
https://essay.utwente.nl/71706/
Ellis, R., Kurniawan, S. (2000). Increasing the usability of online information for older users: a case study in participatory design. Human Computer Interaction, 263-276. doi: 10.1207/S15327590IJHC1202_6.
Epley, N., Waytz, A., Cacioppo, J. (2007). On seeing human: a three-factor theory of
anthropomorphism. Psychological review, 864-86. doi: 10.1037/0033-295X.114.4.864. Fakun, D. (2009). How to mitigate the significant negative influence of computer anxiety on
Fishbein, M., & Ajzen, I. (1974). Attitudes towards objects as predictors of single and multiple behavioral criteria. Psychological Review, 59–74. doi: 10.1037/h0035872 Følstad, A., Nordheim, C., Bjørkli, C. (2018). What makes users trust a chatbot for customer
service? An exploratory interview study. The Fifth International Conference on Internet Science. 194-208. doi: 10.1007/978-3-030-01437-7_16
Go, E., & Sundar, S. (2019). Humanizing chatbots: 'the effects of visual, identity and conversational cues on humanness perceptions'. Computer in Human Behavior, 304-316.
Gray, H., Gray, K., Wegner, D. (2007). Dimensions of Mind Perception. Science, 619. doi: 10.1126/science.1134475.
Hong, W., Thong, J. (2013) Internet Privacy Concerns: An Integrated Conceptualization and Four Empirical Studies. MIS Quarterly, 275-298.
Kim, H., Huh, J. (2017) Perceived relevance and privacy concern regarding online behavioral advertising (OBA) and their role in consumer responses. Journal of Current Issues & Research in Advertising, 92-105. doi: 10.1080/10641734.2016.1233157
Ischen C., Araujo T., Voorveld H., van Noort G., Smit E. (2020) Privacy concerns in chatbot interactions. Chatbot Research and Design, 34-48.
Izard, C. (1960). Personality similarity and friendship. The Journal of Abnormal and Social Psychology, 47–51. doi: 10.1037/h0040056
Johnson, E., Payne, J. (1985). Effort and accuracy in choice. Management Science, 395-414. doi: 10.1287/mnsc.31.4.395.
Kristine L.N., Christian R. (2005) The influence of the avatar on online perceptions of anthropomorphism, androgyny, credibility, homophily, and attraction. Journal of Computer-Mediated Communication, 153–178. doi:
Kulviwat, S., Bruner, G., Kumar, A., Altobello, S., Clark, T. (2007). Toward a unified theory of consumer acceptance technology. Psychology & Marketing, 1059-1084. doi: 10.1002/mar.20196.
Lee, H.-H., & Chang, E. (2011). Consumer attitudes toward online mass customization: An application of extended technology acceptance model. Journal of Computer-Mediated Communication, 171-200.
Letheren, K., Martin, B. A., & Seung Jin, H. (2017). Effects of personification and anthropomorphic tendency on destination attitude and travel intentions. Tourism Management, 65-75.
Liker, J.K. and Sindi, A.A. (1997) User acceptance of expert systems: A test of the theory of reasoned action. Journal of Engineering and Technology Management, 147-173. Doi: 10.1016/S0923-4748(97)00008-8
Lorenzo, Carlota & Constantinides, Efthymios & Alarcón-del-Amo, María-del-Carmen. (2011). Consumer adoption of social networking sites: Implications for theory and practice. Journal of Research in Interactive Marketing, 170-188. doi:
10.1108/17505931111187794.
Marquié, J., Thon, B., & Baracat, B. (1994). Age influence on attitudes of office workers faced with new computerized technologies: A questionnaire analysis. Applied Ergonomics, 130-142.
Maslowsky, J., Buvinger, E., Keating, D. P., Steinberg, L., & Cauffman, E. (2011). Cost-benefit analysis mediation of the relationship between sensation seeking and risk behavior among adolescents. Personality and Individual Differences, 802–806. doi: 10.1016/j.paid.2011.06.028
McGrath, R. (2018, May 2). How to improve customer service with chatbots. Retrieved from Chatbotsmagazine: https://chatbotsmagazine.com/ill-never-buy-from-them-again-using-chatbots-to-avoid-bad-customer-service-e6a967360244
Möller, N. (2018, October 17). Tien jaar Billie: de belangrijkste succesfactoren van de chatbot van bol.com. Retrieved from Retailtrends:
https://retailtrends.nl/item/54221/tien-jaar-billie-de-belangrijkste-succesfactoren-van-de-chatbot-van-bol-com
Norberg, P., Horne, D., & Horne, D. (2007). The Privacy Paradox: Personal Information Disclosure Intentions versus Behaviors. The Journal of Consumer Affairs, 100-126. Nysveen, H., Pedersen, P. E., & Thorbjørnsen , H. (2005). Intentions to Use Mobile Services:
Antecedents and Cross-Service Comparisons. Journal of the Academy of Marketing Science, 330-346.
Pamela J.H., Teresa L.R., Hank J. (2004) Whose Job Is It Anyway? A Study of Human-Robot Interaction in a Collaborative Task. Human–Computer Interaction, 151-181. doi: 10.1080/07370024.2004.9667343
Park, E., Baek, S., Ohm, J., & Chang, H. J. (2014). Determinants of player acceptance of mobile social network games: An application of extended technology acceptance model. Telematics and Informatics, 3-15.
Phelps, J., Nowak, G., and Ferrell, E. 2000. "Privacy concerns and consumer willingness to provide personal information," Journal of Public Policy and Marketing, 27-41. Doi: 10.1509/jppm.19.1.27.16941
Qadir, I., Majid, A. (2019) Mobile conversational commerce: millennial’s perception and attitude towards chatbot adoption in DKI Jakarta. Sekolah Tinggi
Ramachandran, A. (2019). User adoption of chatbots." Available at SSRN 3406997.
Rice, R. E., & Aydin, C. (1991). Attitudes toward new organizational technology: Network proximity as a mechanism for social information processing. Administrative Science Quarterly, 36(2), 219–244. doi: 10.2307/2393354
Sarathy, R., & Robertson, C. J. (2003). Strategic and Ethical Considerations in Managing Digital Privacy. Journal of Business Ethics, 111-126.
Schierz, P., Schilke, O., Wirtz, B. (2010). Understanding consumer acceptance of mobile payment services: An empirical analysis. Electronic Commerce Research and Applications, 209-216. doi: 10.1016/j.elerap.2009.07.005.
Seeger, A., Pfeiffer, J., Heinzl, A. (2017). "When do we need a human? Anthropomorphic design and trustworthiness of conversational agents". Sixteenth Annual Pre-ICIS Workshop on HCI Research in MIS, Seoul. http://aisel.aisnet.org/sighci2017 Shawar, B.A., & Atwell, E. (2007). Chatbots: are they really useful? LDV Forum, 29-49. Shibchurn, J., Yan, X. (2015). Information disclosure on social networking sites: An
intrinsic–extrinsic motivation perspective. Computers in Human Behavior, 103-117. doi: 10.1016/j.chb.2014.10.059.
Spiekermann, S., Grossklags, J., Berendt, B. (2001). Stated privacy preferences versus actual behaviour in EC environments: a reality check, e-Finance, 129-147. doi: 10.1007/978-3-642-59504-2_8.
Steven M. Elias, William L. Smith & Chet E. Barney (2012) Age as a moderator of attitude towards technology in the workplace: work motivation and overall job satisfaction, Behaviour & Information Technology, 453-467, doi: 10.1080/0144929X.2010.513419 Swanner, N. (2016, March 1). Data shows demand for developers who know swift has
skyrocketed. Retrieved from Thenextweb: https://thenextweb.com/dd/2016/03/01/data-shows-demand-for-developers-who-know-swift-has-skyrocketed/
Tae, H.B., Mariko, M. (2012) Stay away from me, Journal of Advertising, 59-76. doi: 10.2753/JOA0091-3367410105
Taylor, J.F., Ferguson, J. and Ellen, P.S. (2015), "From trait to state: understanding privacy concerns", Journal of Consumer Marketing, Vol. 32 No. 2, pp. 99-112. doi:
10.1108/JCM-07-2014-1078
Venkatesh, V., Morris, M., Davis, G., Davis, F (2003). User Acceptance of Information Technology: Toward a Unified View, MIS Quarterly. 425-478. doi:
10.2307/30036540.
Verena M. Wottrich, Peeter W.J. Verlegh & Edith G. Smit (2017) The role of customization, brand trust, and privacy concerns in advergaming. International Journal of
Advertising, 60-81. doi: 10.1080/02650487.2016.1186951
Viergever, A. (2017, May 16). Nederlanders en chatbots. Retrieved from Kayak: https://www.kayak.nl/news/nederlanders-en-chatbots/
Vijayasarathy, L. R. (2004). Predicting consumer intentions to use on-line shopping: The case for an augmented technology acceptance model. Information & Management, 747-762. doi: 10.1016/j.im.2003.08.011
Wang, Y.-S., Wang, Y.-M., Lin, H.-H., & Tang, T.-I. (2003). Determinants of user
acceptance of Internet Banking: an empirical study. International Journal of Service Industry Management, 501-519. doi: 10.1108/09564230310500192
Waytz, A., Cacioppo, J., Epley, N. (2010). Who sees human?: The stability and importance of individual differences in anthropomorphism. Perspectives on Psychological Science, 219-232. doi: 10.1177/1745691610369336.
Weijters, B., Rangarajan, D., Falk, T., & Schillewaert, N. (2007). Determinants and outcomes of customers' use of self-service technology in a retail setting. Journal of Service Research, 3-21. doi: 10.1177/1094670507302990
Wright, S. A., & Xie, G.X. (2019). Perceived privacy violation: exploring the malleability of privacy expectations. Journal of Business Ethics, 123-140. doi: 10.1007/s10551-017-3553-z
Xu, H., Dinev, T., Smith, H.J., Hart, P. (2008). Examining the formation of individual's privacy concerns: toward an integrative view (2008). ICIS 2008 Proceedings. 6. Zaphiris, P., & Ioannou, A. (2018). Learning and Collaboration Technologies. Learning and
Teaching. Las Vegas: Springer.
Zarouali, B., Van den Broeck, E., Walrave, M., & Poels, K. (2018). Predicting Consumer Responses to a Chatbot on Facebook. Cyberpsychology, Behavior, and Social Networking, 491-497. doi: 10.1089/cyber.2017.0518
Zukowski, T., Brown I. (2007). Examining the influence of demographic factors on internet users' information privacy concerns. ACM International Conference Proceeding Series. 226. 197-204. doi: 10.1145/1292491.1292514.
Appendix A: survey
First of all, thank you for taking part in this study!
This survey was developed for my master thesis research project at the Graduate School of Communication, a part of the University of Amsterdam.
In the following survey, a few questions regarding chatbots will be asked. The study will take about 5 minutes.
I would like to inform you that this research is being carried out under the responsibility of the ASCoR, University of Amsterdam, and we can guarantee that:
1) Your anonymity will be safeguarded, and that your personal information will not be passed on to third parties under any conditions, unless you first give your permission.
2) You can refuse to participate in the research or withdraw without having to give a reason for doing so. You also have up to 24 hours after participating to withdraw your permission to allow your answers or data to be used in the research.
3) Participating in the research will not entail you being subjected to any risk or discomfort, the researchers will not deliberately mislead you, and you will not be exposed to any
explicitly offensive material.
4) No later than two months after the conclusion of the research, I will be able to provide you with a research report that explains the results of this research.
For more information about the research, you are welcome to contact me (Eva Teeuwissen) at any time.
Should you have any complaints or comments about the course of the research and the
procedures it involves as a consequence of your participation in this research, you can contact the designated member of the Ethics Committee representing ASCoR, at the following
address:
ASCoR Secretariat, Ethics Committee, University of Amsterdam, Postbus 15793, 1001 NG Amsterdam; +31 020‐525 3680; ascor‐secr‐fmg@uva.nl. Any complaints or comments will
be treated with the strictest confidence. I hope that I have provided you with sufficient information.
I would like to take this opportunity to thank you in advance for your assistance with this research, which I greatly appreciate!
Best,
Eva Teeuwissen
Eva.teeuwissen@gmail.com
Before you begin, I would like you to sign the following informed consent:
I hereby declare that I have been informed in a clear manner about the nature and method of the research.
I agree, fully and voluntarily, to participate in this research study.
With this, I retain the right to withdraw my consent, without having to give a reason for doing so.
I am aware that I may halt my participation in the experiment at any time.
If my research results are used in scientific publications or are made public in another way, this will be done in such a way that my anonymity is completely safeguarded.
My personal data will not be passed on to third parties without my expressed permission.
What is your age in years?
- Male - Female - Other
What is your nationality?
What is the highest level of education you have obtained? - None - Elementary school - Secondary educaiton - Bachelor’s degree - Master’s degree - Doctorate degree - Other (please specify)
What customer service chatbot have you used before? (for example, customer service chatbot from Lufthansa, Nuon, Coolblue, Bol, Ikea) please name the company of the (customer service) chatbot you have used before
Please answer the following questions with regard to the customer service chatbot you just named in the previous question. For example, I have used Coolblue's chatbot before, then I will answer all the following questions having in mind Coolblue's chatbot.
How often do you use your aforementioned (customer service) chatbot? - Never
- Less than once per month - One time per month
