Employee voice — the discretionary sharing of ideas, concerns, and suggestions — is a known source of value inside firms. But as organizations adopt AI systems to listen to workers, the question of how that mediation changes voice and its effects becomes pressing. A new preprint on arXiv addresses this by disentangling two pathways: informational (what is conveyed) and affective (how it feels to be heard).
According to the abstract, the paper is motivated by the growing trend of replacing or augmenting human-led listening with AI. The authors aim to separate the informational and affective routes through which AI-mediated voice influences performance. However, the abstract provided is truncated, so the specific findings, methods, and data are not available from the source alone.
Even without results, the framing is significant. It suggests that the value of employee voice depends not only on the content transmitted but also on the emotional experience of the speaker. If AI-mediated listening changes either pathway, it could alter the performance benefits firms expect from voice programs.