Learning from the Dark Web: leveraging conversational agents in the era of hyper-privacy to enhance marketing

“Adomavicius and Tuzhilin (2005).”

“However, emotion and sentiment tracking are generally based on the current interaction (e.g., Yang et al. 2016, Chen et al. 2016) and can be used to personalize interaction with Buff and Ghost consumers alike.”

“…and execute tasks (Sankar & Balakrishnan, 2016; Shum, He, & Li, 2018), in addition to providing more enjoyable experiences (Brandtzaeg & Følstad, 2018) and a sense of companionship (Turkle 2017; Brandtzaeg & Følstad, 2017), CAs can nudge consumers to voluntarily share personal data with companies.”

“…conscious attention to a subset of contextual cues that trigger various scripts from the past (Langer, 1992; Moon, 2000, 2003; Nass & Moon, 2000; Nass, Steuer, & Tauber, 1994)) that are applied mindlessly even when such behaviors seem irrational, inappropriate, or unnecessary (Nass & Moon, 2000).”

“Excessive anthropomorphism can also trigger consumer discomfort (also known as the “ uncanny valley ” concept — see Mori, MacDorman, & Kageki, 2012; van Doorn et al., 2017) leading to decreased favorability toward the CA (Mende et al., 2019).”

“Consumers weigh privacy concerns and related risks against the benefits of information disclosure (e.g., Dinev & Hart, 2006).”

“…of conscious attention to a subset of contextual cues that trigger various scripts from the past (Langer, 1992; Moon, 2000, 2003; Nass & Moon, 2000; Nass, Steuer, & Tauber, 1994)) that are applied mindlessly even when such behaviors seem irrational, inappropriate, or unnecessary (Nass & Moon,…”

“” The control-based definition of privacy is broadly accepted and has been used as the foundation of most information privacy research (Smith et al. 2011).”

“This discrepancy between attitudes and behaviors is what scholars refer to as the “ privacy paradox ” (Acquisti et al., 2015).”

“The sequence approach uses processual browsing activities to infer consumer preferences (Mannila, Toivonen, & Verkamo, 1997; Niu, Yan, Zhang, & Zhang, 2002).”

“Understanding the message involves semantic encoding and intent understanding, that is, understanding the purpose of the message (e.g., Tur & Deng 2011; Vinyals & Le, 2015).”

“…forcing firms to quickly “ race to the bottom, ” given the perception that consumers would have full information about competitive offerings (e.g., Peterson, et al. 2012), and morphing banner advertising (Urban et al., 2013), which are made possible by leveraging user profiling techniques…”

“…the customer service infrastructure, by replacing or augmenting tasks traditionally performed by sales employees (Larivière et al., 2017; Verhagen, van Nes, Feldberg, & van Dolen, 2014) and by providing consumers with successful service encounters (Larivière et al., 2017; van Doorn et al., 2017).”

“This focus on privacy enables Dark Web users to personally control (and thus limit) access to information about themselves (Altman 1975; Westin, 1967).”

“…we discuss here, CAs must first elicit and collect data from the consumers ( data collection ), use this data to estimate their preferences and build profiles ( building profile ) that they will consequently use to tailor the interaction (Mobasher, 2007; Mobasher, Cooley, & Srivastava, 2000).”

“Excessive anthropomorphism can also trigger consumer discomfort (also known as the “ uncanny valley ” concept — see Mori, MacDorman, & Kageki, 2012; van Doorn et al., 2017) leading to decreased favorability toward the CA (Mende et al., 2019).”

“While anthropomorphization can generate positive marketing results (Aggarwal & McGill, 2007), too much of it can lead to negative effects.”

“…show preliminary evidence of this dynamic, whereby consumers weigh privacy concerns and related risks against the benefits of information disclosure, and sometimes end up trading privacy for monetary rewards (e.g., Caudill & Murphy 2000; Hann et al. 2008; Phelps et al. 2000; Xu et al. 2010).”

“Customer privacy is central to this potential problem (Martin & Murphy 2017, Stewart 2017).”

“…an-thropomorphize 5 CAs, as individuals often feel less inhibited when interacting with anthropomorphic computers, sharing private information (Leong & Selinger, 2019; Turkle, 2017), and even developing personal relationships (Moon, 2000) (see Appendix 1, Table 7 for a review of prior studies).”

“…becoming critical components of the customer service infrastructure, by replacing or augmenting tasks traditionally performed by sales employees (Larivière et al., 2017; Verhagen, van Nes, Feldberg, & van Dolen, 2014) and by providing consumers with successful service encounters (Larivière et…”

“7 According to Shum et al. (2018, p. 6), a CA “ may generate responses in attractive styles (e.g., having a sense of humor) that improve user engagement.”

“In examining specific anthropomorphic features, researchers have shown that certain personality characteristics such as friendliness and expertise (Verhagen et al., 2014) as well as embodiment and communication style (Qiu & Benbasat, 2009) influence perceptions of social presence.”

“Social presence resulting from anthropomor-phization has been associated with trust, engagement, and satisfaction (Kumar & Benbasat, 2006; Picard, 1997; Qiu & Benbasat, 2009; Turkle, 2017; Bleier, Harmeling, & Palmatier 2019).”

“To personalize the consumer interaction, one must understand the consumer (Adomavicius & Tuzhilin, 2005; Johar, Mookerjee, & Sarkar, 2014; Shum et al. 2018; Tam & Ho, 2006).”

“The resulting Digital, Social Media, and Mobile (DSMM) ecosystem has been considered as a source of intelligence (Lamberton & Stephen 2016) for observing, analyzing, and predicting behavior (Bucklin & Sismeiro 2003; Chatterjee, Hoffman, & Novak 2003; Montgomery et al. 2004).”

“…(Claycomb & Martin, 2001), since it increases sales ’ leads, customer acquisition and retention (Bojei et al., 2013; Sahni, Wheeler, & Chintagunta, 2018), firm profit, customer satisfaction, and enables the discovery of novel consumer needs and preferences (Arora et al., 2008; Huang & Rust, 2017).”

“…emotions impact the privacy calculus of consumers, with positive affect leading to lowered perceptions of risk (Li, Sarathy, & Xu, 2011), higher intentions to disclose information (Anderson & Agarwal, 2011), and more self-disclosure (Kehr, Kowatsch, Wentzel, & Fleisch, 2015; Yu, Hu, & Cheng, 2015).”

“This complex trade-off between personalization and privacy is known as the “ personalization paradox ” (Aguirre et al. 2016; Bleier & Eisenbeiss 2015).”

“Social presence resulting from anthropomor-phization has been associated with trust, engagement, and satisfaction (Kumar & Benbasat, 2006; Picard, 1997; Qiu & Benbasat, 2009; Turkle, 2017; Bleier, Harmeling, & Palmatier 2019).”

“These social responses occur as a result of conscious attention to a subset of contextual cues that trigger various scripts from the past (Langer, 1992; Moon, 2000, 2003; Nass & Moon, 2000; Nass, Steuer, & Tauber, 1994)) that are applied mindlessly even when such behaviors seem irrational,…”

“…emotions impact the privacy calculus of consumers, with positive affect leading to lowered perceptions of risk (Li, Sarathy, & Xu, 2011), higher intentions to disclose information (Anderson & Agarwal, 2011), and more self-disclosure (Kehr, Kowatsch, Wentzel, & Fleisch, 2015; Yu, Hu, & Cheng, 2015).”

“In examining specific anthropomorphic features, researchers have shown that certain personality characteristics such as friendliness and expertise (Verhagen et al., 2014) as well as embodiment and communication style (Qiu & Benbasat, 2009) influence perceptions of social presence.”

“…Web 1.0, there was a fear that by disseminating information widely (particularly price information), firms would lose their ability to effectively price discriminate across markets, segments, and purchase occasions, essentially collapsing markets to a lowest common denominator (e.g., Burke 1997).”

“…show preliminary evidence of this dynamic, whereby consumers weigh privacy concerns and related risks against the benefits of information disclosure, and sometimes end up trading privacy for monetary rewards (e.g., Caudill & Murphy 2000; Hann et al. 2008; Phelps et al. 2000; Xu et al. 2010).”

“Data collection Two different mechanisms currently inform the collection of data: explicit and implicit methods (Adomavicius, Huang, & Tuzhilin, 2008; Li & Karahanna, 2015; Murthi & Sarkar, 2003).”

“…has predominantly focused on the effects of P2P file sharing (and other piracy behaviors) on price sensitivity (Jain 2008, Sinha & Mandel 2008), substitution effects (Danaher et al. 2010), diffusion (Givon, Mahajan & Muller 1995), and control mechanisms/policy (Sinha, Machado, & Sellman 2010).”

“However, emotion and sentiment tracking are generally based on the current interaction (e.g., Yang et al. 2016, Chen et al. 2016) and can be used to personalize interaction with Buff and Ghost consumers alike.”

“…more accurate identity graphs that allow companies to market products and services more efficiently and effectively by, for example, targeting people with the right content at the right time (McAfee & Brynjolfsson, 2012; Schumann, von Wangenheim, & Groene, 2014; Spangler, Hartzel, & Gal-Or, 2006).”

“…(Claycomb & Martin, 2001), since it increases sales ’ leads, customer acquisition and retention (Bojei et al., 2013; Sahni, Wheeler, & Chintagunta, 2018), firm profit, customer satisfaction, and enables the discovery of novel consumer needs and preferences (Arora et al., 2008; Huang & Rust, 2017).”

“This complex trade-off between personalization and privacy is known as the “ personalization paradox ” (Aguirre et al. 2016; Bleier & Eisenbeiss 2015).”

“…mechanisms used by firms (Claycomb & Martin, 2001), since it increases sales ’ leads, customer acquisition and retention (Bojei et al., 2013; Sahni, Wheeler, & Chintagunta, 2018), firm profit, customer satisfaction, and enables the discovery of novel consumer needs and preferences (Arora et…”

“Empirical evidence (e.g., Wang & Benbasat 2008) suggests that such transparency, i.e., explaining the logic behind a particular thought, decision, or recommendation, engenders trust.”

“Many systems utilize a collection of individual or aggregate consumer facts (e.g., demographic information, favorite product, amount spent in an online store) to represent factual profiles in relational databases (Adomavicius & Tuzhilin, 2001).”

“This research has predominantly focused on the effects of P2P file sharing (and other piracy behaviors) on price sensitivity (Jain 2008, Sinha & Mandel 2008), substitution effects (Danaher et al. 2010), diffusion (Givon, Mahajan & Muller 1995), and control mechanisms/policy (Sinha, Machado, &…”

“The type of approach leveraged by the CAwill heavily depend on the type of data used to build consumer profiles (Li & Karahanna, 2015), and consequently on whether the consumer is Buff or Ghost.”

“As such, CA personality information (e.g., age, gender, etc.) is often incorporated into the process of generating responses (e.g., see Li et al. 2006; Mathews et al. 2015, Shum et al. 2018).”

“…has predominantly focused on the effects of P2P file sharing (and other piracy behaviors) on price sensitivity (Jain 2008, Sinha & Mandel 2008), substitution effects (Danaher et al. 2010), diffusion (Givon, Mahajan & Muller 1995), and control mechanisms/policy (Sinha, Machado, & Sellman 2010).”

“In the past, a distinction has been made (Smith 2001, pp. 1000-1001) be-tween countries that viewed privacy as a human right and passed “ sweeping privacy bills that address all the instances of data collection, use, and sharing (Bennett & Raab 2006; Dholakia & Zwick 2001) ” versus those that…”

“For example, whereas the Surface Web is currently grappling with the high volatility in the value of Bitcoin (the most familiar example of cryptocurrencies), the Dark Web trades in a variety of different cryptocurrencies, some of which are market-specific.”

“Signatures are the data structures used to capture the evolving behavior learned from large streams of simple transactions (Cortes et al., 2000).”

“This research has predominantly focused on the effects of P2P file sharing (and other piracy behaviors) on price sensitivity (Jain 2008, Sinha & Mandel 2008), substitution effects (Danaher et al. 2010), diffusion (Givon, Mahajan & Muller 1995), and control mechanisms/policy (Sinha, Machado, &…”

“…“ race to the bottom, ” given the perception that consumers would have full information about competitive offerings (e.g., Peterson, et al. 2012), and morphing banner advertising (Urban et al., 2013), which are made possible by leveraging user profiling techniques (Trusov, Ma, & Jamal 2016).”

“…& Garnier, 2012; Chakrabarti & Luger, 2015; Gnewuch et al., 2017; Table 6 Research agenda for personalizing consumer interactions through conversational agents in the era of hyper-privacy Jenkins et al., 2007; McTear, Callejas, & Griol, 2016; Schuetzler et al., 2014; Shechtman & Horowitz, 2003).”

“These approaches include (a) content-based , which typically uses the consumers ’ stored profile that includes historical ratings, viewing behavior, and purchases to match preferences; (b) social networks , which leverages the social connections of consumers to match their preferences assuming that people who are frie…”

“However, in the current evolution to Web 3.0, marketers face the very real possibility of losing these advantages if they no longer have the ability to identify consumers and connect them to their previous behaviors (e.g., Deighton 1997).”

“…“ race to the bottom, ” given the perception that consumers would have full information about competitive offerings (e.g., Peterson, et al. 2012), and morphing banner advertising (Urban et al., 2013), which are made possible by leveraging user profiling techniques (Trusov, Ma, & Jamal 2016).”

“…show preliminary evidence of this dynamic, whereby consumers weigh privacy concerns and related risks against the benefits of information disclosure, and sometimes end up trading privacy for monetary rewards (e.g., Caudill & Murphy 2000; Hann et al. 2008; Phelps et al. 2000; Xu et al. 2010).”

“Adomavicius and Tuzhilin (2005).”

“…have explored traditional marketing questions, like the profitability of vendors and residual value to consumers (Holt, Smirnova, & Chua 2016), while utilizing traditional marketing approaches to guide research in this environment (Li, Chen, & Nunamaker 2016; Benjamin, Valacich, & Chen 2019).”

“Fairness and transparency Alternatively, self-disclosure can be encouraged when firms provide assurances of algorithmic fairness and transparency (Garfinkel et al., 2017).”

“…relationship-building mechanisms used by firms (Claycomb & Martin, 2001), since it increases sales ’ leads, customer acquisition and retention (Bojei et al., 2013; Sahni, Wheeler, & Chintagunta, 2018), firm profit, customer satisfaction, and enables the discovery of novel consumer needs and…”

“…been many failures and many challenges still remain in designing CAs that provide high quality interactions (Ben Mimoun, Poncin, & Garnier, 2012; Chakrabarti & Luger, 2015; Gnewuch et al., 2017; Table 6 Research agenda for personalizing consumer interactions through conversational agents in the…”

“Similarly, other fields have explored traditional marketing questions, like the profitability of vendors and residual value to consumers (Holt, Smirnova, & Chua 2016), while utilizing traditional marketing approaches to guide research in this environment (Li, Chen, & Nunamaker 2016; Benjamin,…”

“However, even those consumers who are fully aware of the costs, and are concerned about their privacy, often choose to disclose identifiable personal data (Adjerid et al., 2018).”

“For example, some attempts to provide overly humanized agents have created unrealistic consumer expectations that turned into frustration (Knijnenburg & Willemsen, 2016) and abuse (Neff & Nagy, 2016).”

“…methods of data collection, session context modelling by CAs (e.g., Shum et al. 2018) allows them to gather and use interactive click stream data during the specific interaction between the CA and the consumer (i.e., implicit data, see Johar et al., 2014; Padmanabhan, Zheng, & Kimbrough, 2001).”

“Other criminology scholars have examined the structure of illicit digital markets, and the associated efficiency and resiliency they exhibit (Bakken, Moeller & Sandberg 2018; Duxbury & Haynie 2019).”

“…& Garnier, 2012; Chakrabarti & Luger, 2015; Gnewuch et al., 2017; Table 6 Research agenda for personalizing consumer interactions through conversational agents in the era of hyper-privacy Jenkins et al., 2007; McTear, Callejas, & Griol, 2016; Schuetzler et al., 2014; Shechtman & Horowitz, 2003).”

“Users can utilize “ anonymity-granting technologies ” to protect their privacy from government agencies, political opponents, trolls, data-hungry organizations, and even Internet service providers (Jardine, 2018).”

“…an-thropomorphize 5 CAs, as individuals often feel less inhibited when interacting with anthropomorphic computers, sharing private information (Leong & Selinger, 2019; Turkle, 2017), and even developing personal relationships (Moon, 2000) (see Appendix 1, Table 7 for a review of prior…”

“For instance, within criminology, researchers have examined the effect of specific news on illicit sales volumes, finding counterintuitive increases in Dark Web transactions (Ladegaard 2019).”

“In a related IS study, Yue, Wang, & Hui (2019) conducted a user-generated content analysis of Dark Web hacker communities, and found evidence connecting increases in user chatter to a lower frequency of cyber-attacks.”

“Other criminology scholars have examined the structure of illicit digital markets, and the associated efficiency and resiliency they exhibit (Bakken, Moeller & Sandberg 2018; Duxbury & Haynie 2019).”

“These techniques may include descriptive models, such as rules, sequences, and signatures (Tuzhilin, 2008), or predictive models such as logistic regressions, neural networks, decision trees, support vector machines (SVM), and Bayesian networks (Adomavicius et al., 2008; Murthi & Sarkar, 2003).”

“To personalize the consumer interaction, one must understand the consumer (Adomavicius & Tuzhilin, 2005; Johar, Mookerjee, & Sarkar, 2014; Shum et al. 2018; Tam & Ho, 2006).”

“As such, CA personality information (e.g., age, gender, etc.) is often incorporated into the process of generating responses (e.g., see Li et al. 2006; Mathews et al. 2015, Shum et al. 2018).”

“Understanding the message involves semantic encoding and intent understanding, that is, understanding the purpose of the message (e.g., Tur & Deng 2011; Vinyals & Le, 2015).”

“Customer privacy is central to this potential problem (Martin & Murphy 2017, Stewart 2017).”

“The sequence approach uses processual browsing activities to infer consumer preferences (Mannila, Toivonen, & Verkamo, 1997; Niu, Yan, Zhang, & Zhang, 2002).”

“Thomaz et al. (2020) identify two types of consumers who vary in willingness to share information.”

“Humanoid service robots and artificial intelligence bots, for example, are increasingly attracting industry attention (Araujo, 2018; Go & Sundar, 2019; Larivi (cid:1) ere et al., 2017; Thomaz et al., 2020).”

“In their paper, Thomaz et al. (2020) argue that a web-wide, hyper-private future may come to mirror practices currently observed in the Dark Web.”