In the real world, we often fail to notice changes in our environment. In some cases, such as not noticing a car moving into our lane, the results can be catastrophic. This so-called change blindness has been seen experimentally both through failing to notice changes to images on-screen as well as failing to notice a change in other people's identity. However, less is known regarding how change blindness manifests in virtual settings varying in visual clutter or with varying types of interaction with someone prior to the change. Across two studies (n = 134), participants engaged in an online video chat with a confederate, with two levels of visual clutter (none, a lot) and three levels of interaction (none, light conversations about weather/TV, deeper conversations about goals/greatest regrets). We found no modulation of change blindness rates across perceptual clutter. Curiously, we found a large discrepancy in change blindness rates in Experiment 1 (79%; 52/66) versus Experiment 2 (16%; 11/68) that we explored, leading to some evidence that increasing the level of interaction led to greater change blindness rates, but only for pairs who identified as belonging to different ethnicities. Taken together, our work suggests that we may pay attention to people differently in virtual settings compared to in-person, that in-group and out-group biases may have an effect on change blindness rates, and that while clutter does not seem to affect change blindness rates, one's level of interaction just might. (PsycInfo Database Record (c) 2024 APA, all rights reserved).