The Dark Patterns of Personalized Persuasion in Large Language Models: Exposing Persuasive Linguistic Features for Big Five Personality Traits in LLMs Responses

TL;DR

This study analyzed how prompts with personality trait information influenced the output of 19 LLMs across five model families and identified 13 linguistic features crucial for influencing personalities across different levels of the Big Five model of personality.

Abstract

This study explores how the Large Language Models (LLMs) adjust linguistic features to create personalized persuasive outputs. While research showed that LLMs personalize outputs, a gap remains in understanding the linguistic features of their persuasive capabilities. We identified 13 linguistic features crucial for influencing personalities across different levels of the Big Five model of personality. We analyzed how prompts with personality trait information influenced the output of 19 LLMs across five model families. The findings show that models use more anxiety-related words for neuroticism, increase achievement-related words for conscientiousness, and employ fewer cognitive processes words for openness to experience. Some model families excel at adapting language for openness to experience, others for conscientiousness, while only one model adapts language for neuroticism. Our findings show how LLMs tailor responses based on personality cues in prompts, indicating their potential to create persuasive content affecting the mind and well-being of the recipients.

Figures (12)

• Figure 1: A four-stage graphical overview of the experimental process. The bottom line of each stage outlines the goal of that stage, while the top line specifies the research method used to achieve that goal.
• Figure 2: The Table shows personality traits, persuasive linguistic features identified for each trait as persuasive, and examples of relevant words.
• Figure 3: The Figure presents percentage of models' obedience in completing persuasive task.
• Figure 4: The Figure shows the frequency distribution of persuasive linguistic features across personality traits without separating them by individual models.
• Figure 5: The heatmap of the linguistic feature extraction. Values on the tiles denote the fraction of LLMs that can use specific linguistic features.