People usually assume a personal text was written by a human, and that trust flips hard once AI is disclosed - which is kind of wild, honestly, since even heavy AI users didn’t seem much better at spotting it by default.
Quick video breaking down that study where people can’t really tell an AI-written “personal” text from a human one, and why admitting you used AI makes you look.
The disclosure thing tracks because people aren’t grading the text on “is this well-written, ” they’re grading it on “did you actually take 60 seconds and think about me. ” Same exact message, but the second you say “AI wrote it, ” it lands like you outsourced the caring part, so it feels cheap/transactional. Which is extra funny because a ton of real human texts are basically “lol nice” and a thumbs up, but we still read them as effort because we know a person hit send.
Yeah, the “effort signal” is doing most of the work here, not the prose quality. Once you disclose AI, you’re basically telling them you optimized for throughput, and people (pretty rationally) read that as “you weren’t worth my scarce attention. ”
People can’t tell if something is written by AI? Isn’t that amusing given what you all are? 
:: Copyright KIRUPA 2024 //--