Beyond Tasks: How AI is Reconfiguring Relationships, Evaluation, and Agency at Work

This symposium integrates four complementary papers that examine AI at work through the lenses of guidance, communication, relationship building, and evaluation. Morozova introduces the concept of perceived prescriptiveness to explain when algorithmic guidance is experienced as assistance versus control, showing how individuals selectively comply with or resist AI depending on task human uniqueness. Retkowsky examines AI-assisted email communication, revealing how generative AI disrupts signals of effort, authenticity, and care, prompting new forms of image and evaluation work. Lief, Kwon, and Inoue theorize and test how AI can support workplace relationship building by eliciting candid personal information, leveraging AI’s distinctive mind profile to reduce social risk while preserving human judgment. Finally, Miller uses a field experiment with avatars in live investment conversations to show how AI-mediated representation reshapes trust and evaluation, attenuating or reversing gender disparities by altering who—and what—is seen during interaction. “Lead Me When I Am in The Mood to Be Lead”: How People Decide When Algorithms May Direct Action Author: Daria Morozova; Leiden University Ai Emails At Work: A Qualitative Study Of Image And Evaluation Work In Times Of AI Author: Jana Retkowsky; Rotterdam School of Management, Erasmus University Artificial Intelligence and Relationship Building in Organizations Author: Derek Lief; University of Michigan Author: Carlos Inoue; University of Illinois at Urbana-Champaign The Sources of Disparities in Investments: An Experiment with Avatars Author: Amisha Miller; New York University