Characterizing and modeling harms from interactions with design patterns in AI interfaces

TL;DR

This work addresses the gap that AI-interface design poses for harms and risk assessment by presenting DECAI, a design-augmented control framework to study repeated human-AI interactions. DECAI integrates a scoping review of harmful AI interface patterns with control-systems theory to decompose how interface features—through affordances and feedback loops—affect user states and welfare, and to generate testable hypotheses. The authors identify four high-signal design-pattern themes (traditional dark patterns, anthropomorphism, explainability/transparency, and seamless design) and demonstrate DECAI via two case studies: recommendation feeds and conversational LLM interfaces. They argue that this framework supports iterative, design-focused impact assessment, informs regulatory considerations, and can be extended to non-adaptive AI and physical interfaces while addressing disparate impacts across user groups.

Abstract

The proliferation of applications using artificial intelligence (AI) systems has led to a growing number of users interacting with these systems through sophisticated interfaces. Human-computer interaction research has long shown that interfaces shape both user behavior and user perception of technical capabilities and risks. Yet, practitioners and researchers evaluating the social and ethical risks of AI systems tend to overlook the impact of anthropomorphic, deceptive, and immersive interfaces on human-AI interactions. Here, we argue that design features of interfaces with adaptive AI systems can have cascading impacts, driven by feedback loops, which extend beyond those previously considered. We first conduct a scoping review of AI interface designs and their negative impact to extract salient themes of potentially harmful design patterns in AI interfaces. Then, we propose Design-Enhanced Control of AI systems (DECAI), a conceptual model to structure and facilitate impact assessments of AI interface designs. DECAI draws on principles from control systems theory -- a theory for the analysis and design of dynamic physical systems -- to dissect the role of the interface in human-AI systems. Through two case studies on recommendation systems and conversational language model systems, we show how DECAI can be used to evaluate AI interface designs.

Figures (1)

• Figure 1: DECAI model consisting of an AI block, user block, and interface block.