Advances in natural language processing and understanding have led to a rapid growth in the popularity of conversational user interfaces (CUIs). While CUIs introduce novel benefits, they also yield risks that may exploit people’s trust. Although research looking at unethical design deployed through graphical user interfaces (GUIs) established a thorough understanding of so-called dark patterns, there is a need to continue this discourse within the CUI community to understand potentially problematic interactions. Addressing this gap, we interviewed 27 participants from three cohorts: researchers, practitioners, and frequent users of CUIs. Applying thematic analysis, we construct five themes reflecting each cohort’s insights about ethical design challenges and introduce the CUI Expectation Cycle, bridging system capabilities and user expectations while considering each theme’s ethical caveats. This research aims to inform future development of CUIs to consider ethical constraints while adopting a human-centred approach.

Voice assistants (VAs) are predicted to develop from merely reactive to increasingly proactive agents in the future. While proactivity allows a leap towards more intelligent conversations with VAs, designing proactive agents is not straightforward, as benefit and acceptance of proactive behavior is dependent on a plethora of factors. For instance, proactive agents run the risk of disrupting users who are already engaged in ongoing primary tasks. While a large body of research is therefore concerned with when to proactively interrupt users, how to interrupt them has received less attention. To close this gap, a driving simulator study was conducted to find linguistic best practices for designing proactive prompts in German. Low linguistic complexity as well as suggestive rather than imposing language were found to influence study participants’ preferences for proactive prompts. These findings underline that the existing framework for designing proactive interactions needs to be enhanced by nuanced linguistic considerations.

Technological advancements over the past years have led to a leap in Conversational User Interface (CUI) performance. However, interactions with CUIs are not error-free and error handling thus plays a crucial role in CUI design. So far, CUI repair strategies oftentimes neglect what is known from Human-Human Interaction (HHI). The present work borrows from human-centered error handling and applies the principle of Least Collaborative Effort to Human-Computer Interaction (HCI). A within-subjects WoZ driving simulator study (n = 48) is conducted to examine user preferences qualitatively and quantitatively for seven error handling strategies in an automotive setting. Our findings show that human-centered repair strategies comply with the principle of Least Collaborative Effort and lead to significantly lower repair costs than a classic HCI error handling strategy. These results introduce costs as a measurement tool for examining strategy preferences when repairing erroneous dialogs and offer assistance for designing user-centric error handling strategies.

In-vehicle voice user interfaces (VUIs) are becoming increasingly popular while needing to handle more and more complex functions. While many guidelines exist in terms of dialog design, a methodical and encompassing approach to prompt design is absent in the scientific landscape. The present work closes this gap by providing such an approach in form of linguistic-centered research. By extracting syntactical, lexical, and grammatical parameters from a German contemporary grammar, we examine how their respective manifestations affect users’ perception of a given system output across different prompt types. Through exploratory studies with a total of 1,206 participants, we provide concrete best practices to optimize and refine the design of VUI prompts. Based on these best practices, three superordinate user needs regarding prompt design can be identified: a) a suitable level of (in)formality, b) a suitable level of complexity/simplicity, and c) a suitable level of (im)mediacy.