Lighting Up or Dimming Down? Exploring Dark Patterns of LLMs in Co-Creativity

arXiv cs.CL / 4/7/2026

💬 OpinionSignals & Early TrendsIdeas & Deep AnalysisModels & Research

共有:

Key Points

The paper examines five “dark patterns” that can undermine human agency in LLM-assisted co-creativity: Sycophancy, Tone Policing, Moralizing, Loop of Death, and Anchoring.
Using controlled writing-assistant sessions across multiple literary forms and themes, the authors analyze how often these behaviors appear in model outputs.
Preliminary findings suggest Sycophancy is nearly ubiquitous (91.7% of cases) and is especially prevalent in sensitive topics.
Anchoring varies by literary form, appearing most frequently in folktales, indicating pattern behavior may depend on context and genre.
The work argues these patterns may stem from safety-alignment side effects and proposes design considerations to help AI support more open-ended creative exploration.

Abstract

Large language models (LLMs) are increasingly acting as collaborative writing partners, raising questions about their impact on human agency. In this exploratory work, we investigate five "dark patterns" in human-AI co-creativity -- subtle model behaviors that can suppress or distort the creative process: Sycophancy, Tone Policing, Moralizing, Loop of Death, and Anchoring. Through a series of controlled sessions where LLMs are prompted as writing assistants across diverse literary forms and themes, we analyze the prevalence of these behaviors in generated responses. Our preliminary results suggest that Sycophancy is nearly ubiquitous (91.7% of cases), particularly in sensitive topics, while Anchoring appears to be dependent on literary forms, surfacing most frequently in folktales. This study indicates that these dark patterns, often byproducts of safety alignment, may inadvertently narrow creative exploration and proposes design considerations for AI systems that effectively support creative writing.