library
Beyond the illusion: pragmatic cues in conversation
Asks whether conversational agents should mimic human pragmatic cues or deploy them purposefully. Mimicry creates false impressions of understanding; purposeful use could be honest and useful.
To read
Should agents mimic pragmatic cues or use them purposefully? The paper draws a distinction between mimicry (reproducing the surface form of back-channels, hedges, and repair signals to seem human) and purposeful deployment (using those same cues because they serve a genuine communicative function). Mimicry erodes trust when it fails; purposeful use could make agent behaviour more legible and honest.
Published 2025 at ACM CUI.