Create one other picture like that one with 5 variations.
It could then generate the second acceptable picture.
As well as, I used to be capable of finding an enchancment over their outdated model.
I’ve used “Hidden Object” pictures previously, the place a number of objects are “hidden” in a picture. They’re enjoyable vocabulary reinforcement actions in my ELL lessons.
Beforehand, Gemini would inform that, although it was accustomed to them, it was unable to generate them.
This time round, although, I gave it this immediate:
Create a hidden object picture of a kitchen, and have the hidden objects all objects that may be sometimes present in a classroom.
It created this picture, and listed these objects: