What agent traits do workers prefer for meaningless tasks?

Within the same 202-worker, 171-task survey, participants also rated 12 agreement-based items describing paired, polarized AI agent traits — such as "Practical vs Imaginative" or "Polite vs Straightforward." The authors report that on tasks already scored high in "bullshitness," "preferences consistently favor efficiency-oriented behavior, prioritizing speed, simplicity, and minimal interaction," and conclude plainly "for highly bullshit tasks, agent design should emphasize constrained, efficient execution rather than explainability or creativity."

The authors tie this to the same mechanism as the delegation-desire finding: because workers already perceive these tasks as disconnected from meaningful outcomes, they don't want an agent to linger on them with transparency, flexibility, or exploration — they want them dispatched. They frame this as a design signal, not just a delegation signal: task-level perceived meaning, they write, "can inform early design choices, including the use of foundation models, tools, or hybrid approaches."

This complicates What UX principles do workplace users want in AI agents?: control, reliability, and context-awareness, weighted there as broad priorities for workplace agents, give way to speed and simplicity once the specific task is perceived as meaningless — the preferred agent behavior is task-meaning-dependent rather than a fixed ranking that holds across all agent work. It extends Does perceived task meaninglessness drive worker desire for AI delegation? from the question of whether workers want a task delegated to the question of how they want the agent to behave once it is.

The excerpt does not establish this as tested behavior: the 12 trait items are stated preferences on agreement-based pairs, not observed outcomes from workers actually using an efficiency-tuned versus exploratory agent, so the finding is what workers say they'd want, not evidence that constrained execution performs better or is experienced better in practice. The sample is the same US-only, self-report group as the delegation finding, and the authors do not report whether this preference holds for tasks that are ambiguous rather than cleanly rated meaningless or meaningful.