This seems like a know your audience situation.
>However, in a world of commodity AI models, can simply use a model from an AI provider that does not watermark. Or download any open source model and run it themselves
That's seems above the skill of the majority of the population that might be tempted at using generative text. I'm thinking of students trying to "write" a paper, or any of the myriad of other things people are blissfully unaware of how genAI works that are using it every day. For those people, that just use the default prompt to blindly copy/paste away. For techy nerd types running from CLI, yeah, they could simply use a different model.
Don't underestimate the whisper network of children. It only takes one to realize that Claude gets caught every time, but super paper writer.example.com flies under the teacher's radar.
With a strong enough regulatory framework, we probably could suppress unlisenced providers and non-watermarking local harnesses to the point where bypassing them would take some technical know-how. However, I do not think there is anywhere near the political will for such a regime.