How AI-Generated Language Shapes User Trust in Social Media? A Systematic Review

Abstract

The rapid development of Generative Artificial Intelligence (GenAI) has transformed social media communication by enabling AI systems to produce increasingly human-like language and content. Although research on AI-generated communication and user trust has expanded, findings remain fragmented across platforms, communication contexts, and linguistic characteristics. This study systematically synthesizes empirical evidence on the relationship between AI-generated language and user trust in social media communication. A Systematic Literature Review (SLR) was conducted following the PRISMA 2020 guidelines. Literature was retrieved from the Dimensions database in July 2026 using predefined search terms related to Generative AI, language, trust, and social media. Of 205 records initially identified, 54 full-text studies were assessed for eligibility, and 23 empirical studies were included in the final synthesis. The findings show that research is concentrated in recent years, particularly in marketing and digital communication, health communication, and misinformation and synthetic media. Across the studies, trust is associated with recurring linguistic and communicative features, including human-likeness and conversational style, emotional expression and empathy, personalization, formality and readability, narrative and experiential language, semantic and informational characteristics, transparency and disclosure, and explanatory language. The findings indicate that user trust is not determined by AI authorship alone but emerges through the interaction of relational and epistemic cues within specific communication contexts. This review contributes to human–AI communication research by identifying linguistic and communicative mechanisms through which AI-generated social media content shapes perceptions of trust, credibility, and authenticity.