Abstract
Imundo et al. (2024) argue that (1) ChatGPT and other chatbots powered by generative Artificial Intelligence (genAI) models can enhance experts' cognition and training, and (2) these benefits are not equally distributed across all users. This commentary commends both arguments. I provide empirical evidence from the field of Human-Computer Interaction (HCI) to show that both statements hold true, not only with ChatGPT, but also with older computational technologies that mimic the behavioral expressions of expertise. By reinforcing Imundo et al. 's arguments, I also argue that now is an opportune time to pursue an even deeper understanding of the relation between ChatGPT and expertise. For instance, how might systems like ChatGPT influence or redefine what it means to be an expert? Rather than distinguishing cognitive, social, and physical expertise, could alternative taxonomies like generative versus evaluative abilities offer more insightful ways to study genAI's cognitive impact?