Here, I examine a hidden challenge of redesigning work around digital tools: even when technology makes interaction easier, it can weaken trust. Focusing on the growing use of avatars in virtual work, my coauthor and I show that people trust a human-form representation more than an avatar representation of the same person, and that trust drops especially when an interaction shifts from a human face to an unfamiliar avatar. The reason is not simply the technology itself, but how people perceive it: avatars can feel less familiar and less similar to us, which dampens trust. This matters for a wide range of organizational activities (e.g., remote onboarding, training, and customer interaction) as these settings increasingly depend on digital self-representations. The broader message of the paper is that, as organizations go for virtual and AI-mediated work, they need to think not only about efficiency, but also about how to prevent the use of digital technology from weakening the social glue of collaboration—trust.

With my coauthors, I explore how people and generative AI can create ideas together without letting human creativity wither away. Much current thinking assumes that the best way to work with AI is to divide labor: let AI generate options and let humans judge them. we push back on that logic by showing that, in creative work, this approach may gradually weaken the very human capacity we care about preserving. Instead, we propose a different model in which humans and AI both stay actively involved in idea generation, and then combine their ideas. We show that this can be powerful because people and AI search for novelty in different ways: AI can explore large conceptual spaces quickly, while humans bring lived experience, context, and intuition that machines do not have. But we also highlight an important trade-off: the more we rely on AI to combine ideas, the more novelty we may gain; yet, the more humans stay involved in combining ideas, the more we preserve human creativity.