A natural portrait of a Japanese woman in a beige shirt, seated by a window in a wooden interior with soft daylight

Prompt

At first, I was generating images using only instructions like “a beautiful Japanese woman,” trying to create a girl who looked cute and realistic, but
no matter what I did, it still retained that AI look, like in the first image....

So I decided to test what would happen if I reworked the prompt with a subject depiction that AI could understand more easily, rather than focusing on the overall mood!

可愛くてリアルにいそうな子を作ろうとして最初は「綺麗な日本人女性」みたいな指示だけで生成してたけど、
どうしても１枚目みたいにAI感が残っちゃう....

そこで、雰囲気じゃなく AIが理解しやすい被写体像を
意識してプロンプトを組み直したらどうなるか検証してみた！