Copy link to clipboard
Copied
I have been using AI in BETA for a while now and a basic upgrade has now moved it to the main stream PS.
In BETA I have had good success in "touching up" photos , removing unwanted features etc. However Now it is in main stream I have come across two issues which did not seem to occur in BETA.
1) When removing a selection eg an unwanted person, instead of mimicing the background I have foung that AI is adding random features, eg. another AI generated person or a fictional advertising board etc.
2) If my subject is holding, again as an example, a handbag which I want to remove, I get a message saying that My request is breaking the "rules".
I would emphasis that none of my photos are sexual in anyway so I am at a loss to know why what was working perfectly in the majority of cases is now producing enexpected results.
Is anyone else experiencing these problems or have an explanation. I have sent the files to PS as examples.
Hi @FWG99 what kind of prompts are you using when running Gen Fill? As noted quite heavily a lot of false flag violations are easily bypassed by entering a period ( . ) in the field first.
Copy link to clipboard
Copied
gen fill might be out of the beta Ps, but it's still very much a beta technology. Expect "bugs" and many of them.
Copy link to clipboard
Copied
Hi @FWG99 what kind of prompts are you using when running Gen Fill? As noted quite heavily a lot of false flag violations are easily bypassed by entering a period ( . ) in the field first.