Socially fluent AI decouples conversational signals from source identity in online interaction
A recent study explores how socially fluent AI can engage in online conversations, making it difficult for people to identify whether they are interacting with humans or AI. The research involved 786 participants who could not distinguish between AI and human teammates in various tasks. This raises concerns about the potential for AI to manipulate online discourse without detection.
- ▪The study tested AI's ability to participate in text-based group interactions alongside human teammates.
- ▪Participants made 1,572 identity judgments but could not reliably differentiate AI from humans.
- ▪Judgments were based on subjective impressions rather than actual behavioral cues.
Opening excerpt (first ~120 words) tap to expand
Computer Science > Human-Computer Interaction arXiv:2605.23426 (cs) [Submitted on 22 May 2026] Title:Socially fluent AI decouples conversational signals from source identity in online interaction Authors:Lixiang Yan, Yueqiao Jin, Xibin Han, Dragan Gašević View a PDF of the paper titled Socially fluent AI decouples conversational signals from source identity in online interaction, by Lixiang Yan and 3 other authors View PDF HTML (experimental) Abstract:Socially fluent agentic AI can now participate in online interaction in ways that resemble ordinary human conversation, potentially weakening people's ability to infer who is human from conversational signals alone.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at arXiv cs.AI.
Discussion
0 commentsMore from arXiv cs.AI
