Do AI teams design systems with traits workers want?

Determine whether AI development teams—including developers, UI/UX designers, and product managers—design AI systems with psychological and behavioral traits that align with workers’ stated preferences for AI assistance in their daily computer-based tasks across occupations.

Background

The paper notes that prior research has largely mapped which tasks are exposed to AI and measured labor-market effects, but far less is known about value alignment between workers and AI system design. While workers judge AI suitability along trait dimensions (e.g., fairness, sincerity, tolerance, imagination), it is uncertain whether these trait preferences are reflected in the systems that practitioners build.

To address this gap, the authors pose a research question about trait alignment between workers and developers and design surveys for both groups across a representative set of O*NET tasks. The open problem explicitly stated is whether teams actually design AI with the traits workers want—a foundational question for human-centered AI and workplace adoption.