I won’t read this — for me it’s just straight-up awful AI.

Yesterday I experienced a cognitive bias.

Context: I built an AI that writes new texts based on my previous ones. And for the first time, I actually liked the result and published it without edits. I felt really inspired. Finally!

I showed it to a marketer I know, and she said: “I won’t read this — for me it’s just awful AI.”

It felt like that dress experiment. I said: “It’s black.” She said: “It’s gold.” I look at it — no, it’s black. She goes: “But all the signs show it’s gold!”

And I started thinking: yes, we perceive content differently. But are there really clear signs of AI that a human couldn’t also use? Are there actually bad texts, or just texts that didn’t work?

The question isn’t whether the text is good or bad. The question is whether identifying a text as AI-generated is really that simple.

I remember there was a time when photographers overused filters and made skin look too smooth. I was among those who preferred a natural look, and it really bothered me when I saw excessive editing. But clients wanted it — they liked it, they admired those filters.

Cognitive biases. How do you deal with this?