PulseExploreJournal ClubResearchersJournals
Instagram
HomeJournal ClubExplore
Synapse
⌘+K
Synapse
May 3, 2026Open Access

Pragmatics In Human-AI Interaction: A Linguistic Study Of Conversational Agents

View Full Paper
Ask AI
Bookmark
Share

Authors

DTDr. S. ThivyanathanDADr. R. Anusha

Discussion

Loading...

Member takes

Overview

Linguistic study examines pragmatic competence in conversational agents, suggesting areas for improvement.

Key Points

  • This research investigates how conversational agents understand and produce contextual meaning in human-AI interactions.
  • Empirical analysis of 50 transcripts of human-AI conversations.
  • Experimental work with 36 participants comparing five conversational agents: ChatGPT-4, Google Bard, Microsoft Copilot, Claude 2, and LLaMA 2.
  • Assessment of rule-based and transformer models' pragmatic capabilities.
  • Transformer models exhibit emergent pragmatic competence, accurately interpreting indirect speech acts in 76% of cases.
  • Gricean implicatures are recognized in only 34% of instances.
  • Cross-turn common ground consistency is observed in only 41% of examples.

Cite This Study

Thivyanathan et al. (2026) studied this question.

synapsesocial.com/papers/69f6e60f8071d4f1bdfc6b6fhttps://doi.org/10.5281/zenodo.19946122
View Full Paper
Ask AI
Bookmark
Share