Today, organizations increasingly utilize conversational agents (CAs), which are smart technologies that converse in a human-to-human interaction style. CAs are very effective in guiding users through digital environments. However, this makes them natural targets for dark patterns, which are user interface design elements that infringe on user autonomy by fostering uninformed decisions. Integrating dark patterns in CAs has tremendous impacts on supposedly free user choices in the digital space. Thus, we conducted a qualitative study consisting of semi-structured interviews with developers to investigate drivers of dark patterns in CAs. Our findings reveal that six drivers for the implementation of dark patterns exist. The technical drivers include heavy guidance of CAs during the conversation and the CAs' data collection potential. Additionally, organizational drivers are assertive stakeholder dominance and time pressure during the development process. Team drivers incorporate a deficient user understanding and an inexperienced team.
Kollmer, Tim; Hauser, Alessa; Oberhofer, Viviana; Blossey, Gregor; and Eckhardt, Andreas, "Uncovering Drivers for the Integration of Dark Patterns in Conversational Agents" (2023). Wirtschaftsinformatik 2023 Proceedings. 6.