What's the state of "invisible watermarks" in generative AI?

I agree they should get over it but in practice our thoughts matter little, and if you pay attention the sentiments don't seem to be changing anytime soon. A lot of people and artists outright harass and spam the accounts of artists/people who are revealed to have something to do with generative AI. They don't care what quality the work is, what you used it for, in what capacity and how long you spent to generate it, they only care about the fact it's AI and want to publicly shame you. That's what I called "the mob."

Also, we already have stuff like Steam labelling your game as "using AI" if it used generation for some assets, for example. How is it different from inspecting whether some elements of your game were ripped off/inspired by another game and putting a "shaming" label on your product, so certain people won't buy it and buy another game instead? (Of course... the latter never happened, it would be considered idiotic, but labeling AI usage is fine). Think. Regardless of why these things are happening, for near-foreseeable future it's creating the world of "better" and "worse" (even if AI was the least significant 5-10% of your work, but you'll still get the label). Same for Amazon and other platforms, I believe, and it's probably going to get worse. Look at Shutterstock - this is what they outright say in their policies against commercial AI content. They don't say they would rather ensure max quality or anything like that, as worry against possible AI-generated trash. Nope. They say you wouldn't be able to own it because it was made (=stolen) by infringing on other IPs, which is a massive misconception. Which is exactly what the mob believes too. But generative AI models don't make collages from real art or anything like that, they barely work like a human would - by analysis and inspiration, as reference to how things work and should be done, no different than real artist being inspired by specific style viewed on Google and incorporating it in their own art.

Anyway, what you use AI for? It's possible it's completely fine with zero issues for some stuff, say, marketing and what not, because nobody cares (wink wink, as long as your image won't get detected and some social media platform/reddit sub decides to ban them). But for any 'recognizable IP' or creative work that deals with social following/consumers, idk, I don't like the idea of getting labelled because X % of the work had AI usage in it. And even then I would have probably worked hard to make the generation perfect and specific, which is not that easy. Not to mention some people use AI to speed up the work and then manually focus on the important parts of their art.

Sorry for the rant, just wanted to clarify what I meant. It's certainly both curious and worrying, but also confusing, what's going on. Especially for someone like me who is not able to inspect the code themselves. My guess is most (if not all) online AI generative tools create some sort of invisible watermark already, even if crude. Some local tools and models might not, but hardly anyone talks about it (with how much noise there is in this space), so not sure. That's why I asked whether somebody knows something more on the matter or maybe even investigated it.