Maybe it already was fraud. I expect to see class-action lawsuits on the way soon.
This is another facet of a larger phenomenon, that I just commented on in another thread[0] - there's a lot of harmful activity that's considered legitimate, thus invisible to our legal and cultural immune systems, yet no less harmful than the slightly less legitimate scams.
--
After reading your post, I had a thought: Making this story about OnlyFans is the click bait element. Seriously, replay the story in your head where it says that top 1/2/3/4/5% of YouTubers or Instagrammers or TikTokkers are using LLMs to answer comments and DMs. Suddenly, the story is much less enticing. It probably would not even make HN front page. Most readers here would look at such an article, shrug their shoulders, and say: "Yeah, seems right. I would do the same.". However, as soon as you add sex work into the mix, it gets way more spicy and elicits more emotional responses.
That said, I don't agree that using AI chatbots to deceive fans is OK anywhere else either. And no, I would definitely not ever do it myself.
I guess there is some truth to the adage that a male body can't push blood both to the brain and to the penis at the same time - it's either one, or the other. I mean, if you step back and take a cold look at the content produced by popular OF performers, it's pretty apparent those are business operations. There are other people involved. Someone does the camera work. Location work. Promotion. And yes, fan management too.
Same applies to YouTube personas, too. Brands, not people. Talking to fans is a job that can be outsourced. AI doesn't introduce anything new; AI features on the platform are competing with creators' brand management agencies, which themselves likely already use ChatGPT for this anyway, because why wouldn't they? ChatGPT does better job at this stuff and is cheaper than human labor - but again, the AI angle isn't important here - that human labor was just a generative model in a protein substrate anyway.