#like it’s okay chatbot you can flatly tell me I’m wrong!

[poorlittleyaoyao]

I went on ChatGPT for the first time today to prove a point to my students about how using AI on their work is academically dishonest and inaccurate, and while I succeeded in my objective (getting it to tell me something objectively false), I was struck by how distressingly polite it is. It used exclamation points to soften anything that could be construed as a command. Most of its replies began with a variant of “I’m sorry.” It apologized to me profusely as it gently and patiently tried to explain to me that horses were not, in fact, part of an automobile.

Anyway, all this is to say that ChatGPT has evidently been trained to use Jin Guangyao’s Koi Tower Customer Service Voice.