Why Muse and Grok Bot ship a blinking mascot
Citing a 2024 study of 580 people, the author argues that human-like assistants are forgiven more for errors, so mascots shift users from judging a calculator to judging a coworker.
most ai products have a lil guy now and apps like muse, grok bot etc are built around a mascot/blob that blinks and emotes at you. there's a 2024 paper that i think explains part of it. it starts from two competing theories of how we judge chatbots: > machine heuristic, where we expect it to be precise and come down hard on it when it isn't > casa (computers are social actors), where we treat it like a person. 580 ppl used chatbots that were more or less human like, and the human like ones got forgiven more for their errors, through two routes. emotionally, the mistakes felt less severe. cognitively, people were more likely to blame themselves, as if they'd asked the question wrong. the paper doesn't test faces specifically, but my read is that the blob works as a switch between those two modes. a plain text box gets judged like a calculator. when you give it eyes and a mood and you start grading it like a coworker, and you'll forgive a coworker for a lot more than you'd forgive a calculator. https://scholars.cityu.edu.hk/en/publications/when-chatbots-make-errors-cognitive-and-affective-pathways-to-und/