Will Ultraforge Engage in Censorship?

I hope nothing related to censorship, copyright or other rules will be applied.

+ 1

Quote

Eric Edelman wrote: While testing the current Photoshop Beta, which introduces the Generative AI Fill feature, my text prompts were censored on two occasions. PS Beta informed me that my prompts "violated Community Standards" and that they were blocked. I was unaware that Adobe had cast itself as a social media entity, implying that a "community" and its "standards" defined what was acceptable speech or expression in Photoshop. My question is: will Ultraforge censor text prompts, on the basis of so-called "Community Standards" or for any other reason?

I have been testing (FireFly) for a little while before the public beta. Some words seem to not be liked by the AI. You could replace or reword your prompt or words in your prompt. One word it does not like is "insane" so I replace it with "crazy" no issues and still has the same effect that I want. Same with proper names such as artists' names. You just have to put it in the right spot for the Ai to not notice. Even numbers it does not like except when put in the right context. Same as things about war etc. I have a prompt which includes stuff about war, which is in the context of card games but firefly would not let me use that until I reworded that.
So IMHO it is not a censorship thing but more of a community guideline thing that was setup to avoid any issues later down the road if one wants to upload an image to their gallery on THEIR site....Sometimes just an overall blanket word ban works or might cause an issue.
If you have been working with AI long one should know that all AIs have different rules that they follow about what words they allow. One prompt will work in DeepDream but that prompt will be too long for Nightcafe or some of the words it does not like. (which I have run into that problem) is that censorship because I have to trim that prompt down or take out words if used in Nightcafe?..... Also all prompts across several AIs will not have the same results of an image.

Quote
Vladimir Golovin wrote: Since UF is a desktop app that runs locally, we are not planning to censor prompts in any way. It's possible that there will be an optional safety checker, but the current plan is to allow users to turn it off.

Thank you very much. I appreciate your straightforward answer.

I understand and accept that certain types of content (violence, explicit sexual activity or non-artistic nudity) will be excluded from being displayed, according to the Terms of Service of a website. However, as you made clear, there is a vast difference between that kind of restriction and interfering with a content creator in the midst of making a composition. The latter case involves censorship of an unacceptable type.