Computers as Bad Social Actors: Dark Patterns and Anti-Patterns in Interfaces that Act Socially

Abstract

Fig. 1. Mock examples of notification messages with manipulative, rude, or tactless framing The Computers Are Social Actors paradigm suggests people exhibit social/anthropomorphic biases in their treatment of technology.Such insights have encouraged the design of interfaces that interact with users in more social (chatty or even friend-like) ways.However, in typical 'dark pattern' fashion, social-emotional responses to systems as (seemingly sentient) agents can be harnessed to manipulate user behaviour. An increasingly common example is app notifications that assume person-like tones to persuade or pressure users into compliance. Regardless of being manipulative, difficulties meeting contextual social expectations can make automated social acting seem rude, invasive, tactless, and even disrespectful-constituting 'social' anti-patterns. This paper explores ways to improve how automated systems treat people in interactions. We mixed four qualitative methods to elicit user experiences and preferences regarding how automated systems "talk" to/at them. We identify an emerging 'social' class of dark and anti-patterns, and propose guidelines for helping (social) interfaces treat users in more respectful, tactful, and autonomy-supportive ways.CCS Concepts: • Human-centered computing → Natural language interfaces; Empirical studies in interaction design; User centered design; Contextual design.

The Voice: Lessons on Trustworthy Conversational Agents from 'Dune'

“…There is a synergistic body of research on "dark patterns"manipulative tricks used to influence user behavior [1,3,5,7,20]. We may be able to incorporate this knowledge to develop our own Truthsayers, or "White Hat AI", to counter manipulative intent.…”

“…Generative imagery and deepfakes now represent a significant concern. They range from the disturbingly convincing videos of "Kari Lake, " produced as a demonstration of this capability by Arizona Agenda, 1 to the amateurish, yet successful deepfakes used to harass Baltimore County Principal Eric Eiswart, which were created using commonly available tools. 2 Importantly, manipulation need not involve wholly fabricated content.…”

Exploring Privacy Implications of Devices as Social Actors

“…An example is Kitkowska et al's [43] 2020 study where they improved privacy notices through designing for a positive emotional effect, but they did not explicitly position their work through a CASA lens. On the other hand, Alberts and Van Kleek's 2023 paper, "Computers are Bad Social Actors: Dark Patterns and Anti-Patterns in Interfaces that Act Socially", shows how certain applications exploit social behaviours to manipulate users into certain actions [6]. They then present guidelines, such as respecting a user's personal boundaries, to avoid interface designs which exploit social behaviours.…”

“…We used Kendall's tau rank correlation coefficient 5 with Holm-Bonferroni adjustment 6 to look for correlations between the means of the device responses and the person responses per theme. According to van den Berg [81], tau > 0.07 signifies a weak correlation, tau > 0.21 is the threshold for a medium correlation, and tau > 0.35 is the threshold for a strong correlation.…”

“…As mentioned previously, trust is an important consideration for privacy design. Another instance is Alberts and Van Kleek [6]'s 2023 study where they used the CASA paradigm to understand deceptive design patterns. They identify a "social-class" of deceptive design pattern and suggest guidelines for how devices should properly "talk" to users [6].…”