Can ChatGPT reciprocate impoliteness? : The Al moral dilemma

Tantucci, Vittorio and Culpeper, Jonathan (2026) Can ChatGPT reciprocate impoliteness? : The Al moral dilemma. Journal of Pragmatics. ISSN 0378-2166 (In Press)

Text (Revised manuscript_latest version_)
Revised_manuscript_latest_version_.docx - Accepted Version
Available under License Creative Commons Attribution.
Download (3MB)

Abstract

Can AI ‘learn’ to be (verbally) violent? Unfortunately, it can. This paper tests ChatGPT 4.0 against real-life impolite interactions to assess whether it responds to human patterns of verbal conflict. Drawing on the Principle of (Im)politeness Reciprocity (Author.b & Author.a, 2021), we prompted ChatGPT 4.0 turn by turn with authentic disputes, tracking its responses through network analysis and Bayesian regression. Our results reveal what we call the ‘AI moral dilemma’ (AI-MD): Large language models are globally aligned to avoid impoliteness through moderation and reinforcement learning, yet locally designed to emulate human conversation—where (Im)politeness Reciprocity is intrinsic. When exposed to sustained impoliteness from real human disputes, the system’s context-sensitive ‘working memory’ can override its moral safeguards, progressively leading it to reciprocate to impolite behaviour. However, this contradiction—between the imperative be moral and the imperative be human-like—marks a fundamental limit for AI pragmatics and raises pressing questions for AI ethics, alignment, and the risks of replicating human conflict. This led to a second finding about AI’s moral dilemma effects: while ChatGPT learns to approximate human patterns of impoliteness reciprocity, it often circumvents outright insults by resorting to implicational impoliteness (cf. Author.b, 2011). This has important implications for AI’s ability to simulate social intentions and Theory of Mind.