Hey anti-AI writers and creative people! Does anyone have advice for dealing with the knowledge that generative AI will probably be trained on any creative work you share outside of a tiny, offline community?
I used to write a lot, but I’ve been struggling to find the motivation to write anything real for a while. I start to get excited thinking about a writing project, and then I remember that whatever I share will most likely be used to build technology that I believe is actively harmful to humanity, and all my motivation drains away.
Several large chunks of my past writing have definitely been sucked into LLMs (multiple long stories about that, not getting into them here). I expect that something similar will eventually happen again if I share my writing in the future. I’ve been looking into ways to prevent LLM training on text, and as far as I can tell, there’s nothing that works well without severely limiting the reach of your writing.
I read and write partially to connect with other people. I’m not a best-selling author or trying to become one, but I want to connect with strangers and isolated people I’ll never meet in real life again, like I did with my past blogs and stories that were later sucked into LLM training. There’s something depressing about severely limiting the potential human reach of my writing (or any other kind of creative work) just to prevent nonconsensual scrapers and book scanners from getting it. That seems like another way of letting the AI industry win. I want to write and share with humans anyway, because those fuckers shouldn’t get to steal my own or anyone else’s joy, but I’m struggling to find my old motivation.
Does anyone have advice for dealing with this issue? Something that’s not just “accept that other people’s actions are out of your control” or “that just comes with the territory nowadays” or “everything in life has risks?” I love writing. I’ve really tried to come around to the “that just comes with the territory nowadays” mindset, but that hasn’t gotten rid of the demotivation. Most of the “writing motivation” advice I’ve found assumes that lack of motivation is rooted in fear of failure, setbacks, getting stuck in the middle of a project, working on the wrong project, etc. I need advice from people who understand this particular problem.
One thing I’ve wondered about: I’m sure the bots that are going around scraping don’t care about the content at all that are just scraping everything they can get. However, I wonder if there’s any analysis done on the content to decide whether it should go into the model or not. Like maybe if extremist or racist or other kinds of sexual content are in it then maybe it would kick it out? If that’s the case maybe such content be included in the page in such a way that it would be invisible? Caveats are of course assistive devices which may pick up the text, and also if somebody dug into it they might make a false claim of racism or whatever.
Unfortunately, that analysis work is done by humans in poorer countries who get paid under $2/hr to sort through horrific content all day. The content might still get scraped, but the problematic parts would probably be de-weighted or left out. All you’d accomplish with that tactic is maybe contributing to an exploited Kenyan worker’s PTSD.
This is what “analysis done on content” looks like: https://time.com/6247678/openai-chatgpt-kenya-workers/
I know the ingested content gets filtered, to discard probably poisoned content. I don’t know the granularity of the filter, but maybe you can do one real paragraph, one markov chain bs paragraph, repeat. That way, the data will probably seem low quality or poisoned to the filter, if it doesn’t treat each paragraph separately.