I don’t think that’s true. The prompt was “images of people” and that is what was delivered. The AI generated something to meet the prompt. That’s all.
You guide the output to influence it.
Like making it “diverse” (include a black person), a person being attractive and free of blemishes unless instructed, a person being a specific age like 20-30 unless instructed.
The prompt wasn’t to generate two equally statistically likely people or something. It was: show me two people. The AI did just that.
Random is implicit in that statement.
I don’t think that’s true. The prompt was “images of people” and that is what was delivered. The AI generated something to meet the prompt. That’s all.
The AI did that. Because of a guided output.
What does “guided output” mean in this context?
You guide the output to influence it.
Like making it “diverse” (include a black person), a person being attractive and free of blemishes unless instructed, a person being a specific age like 20-30 unless instructed.
That are just a few examples what I meant.
That does not appear to be what’s happening in that image.