"AI is allowed as long as it’s used ethically and legally" This is a bit of a paradox, since AI doesn't really have any ethical, or technically legal, use-cases at all. More on that with my second point
"Don’t use outputs that infringe on copyright or trademarks" This is virtually impossible, because AI doesn't pull its information out of thin air or the ether. ANY and EVERY information or data AI has has been trained off of something/somebody else's works. That's how and why AI is able to mimic different artstyles. Of course what's probably meant more with this guideline is "Hey, don't go around creating Disney or Studio Ghibli -prompts" etc. , but I've got the feeling that that's done here only because that'd be the most egregious, blatant and in-your-face thing, and not for any other reason
Of course, you people are free to run your own Jam however you want. I just find the way these guidelines are phrased funny, because if you really think about it, they are oxymorons
I have to agree with your general statement and core issue. If one is using the more "open to use AI". However apps/tools like https://cascadeur.com/ uses AI to help create animations, which picks from their animation library. However the sourcing of that data I don't know, so it could still fall under the Unethically aspect as I can't say if it's all inhouse/mocap sourced, or they've just pulled freely from anyone and everyone.
But it's possible to have something that provides AI-assistance (from my above example) to things we're more commonly referring to when mentioning "AI" which is your point of "they just use whatever they find, no cares given" with a simple prompt