Certainly, authors have managed well without the assistance of AI until now. However, the emergence of new tools brings forth fresh avenues for creativity. I experimented with the 'cut-up method' following my encounter with William Burroughs' work. Jeff Noon's analogy of sound effects such as distortion and echo also inspired novel thoughts in my writing process. Learning about the Oulipo movement led me to explore constrained writing, a practice that remarkably enhanced my creative output. In a similar vein, I regard AI as a tool that can transform our thinking — which is where our words come from. My view may be tangential to the focus of this discussion. I like AI as an aid, rather than for direct content creation. I agree it's problematic in this latter regard.
Writing with AI is akin to giving a highly autistic, demented eight-year-old, high on cocaine exclusive access to your Scrivener. What results is an utter mess. AI just isn't smart enough to write a novel that makes coherent sense from beginning to end. ChatGPT and other LLMs literally have the memory of goldfishes, as in, they can remember around eight thousand words before they forget everything that came before word eight thousand and one.
Unlike the eight year old in your example, we are talking about a computer program, which can save the output to a file, which gives them long term memory. I am not saying they do that, just that the capability is there.
That’s not how any of this works. See this is what I’m talking about when I say that people who don’t understand how LLMs and these data sets work, pushing bandwagons of blatantly false opinions. The capability for something to happen does not compare to what is happening. I could take all of my accumulated ebooks and dump them on the web. Just because I can, doesn’t mean I am.
That's just not true. Author says 'AI-generated' books were published under her name. Amazon wouldn't take them down Self-publishers must declare if content sold on Amazon’s site is AI-generated Focus: ChatGPT launches boom in AI-written e-books on Amazon A Publisher’s Statement Regarding AI-Written Submissions It is true, though, that there is a greater flood of AI generated content in the short story market. Science fiction publishers are being flooded with AI-generated stories AI-generated fiction is flooding literary magazines
I didn't read the articles, but the focus of them seems to be just on the fact that the books were published, not how coherent they might be? Amazon doesn't check books before publishing them to make sure they're coherent. I don't think it checks them for anything. Kindle Publishing is basically just a bot programmed to allow bot-written content to get in.
My point wasn't that the output was being stored, simple that it could be. The disc space requirements for that would be huge. We get that you are biased towards AI, due to the financial interest of your Husband's employment. The debate as I see it, isn't if the technology itself is good or bad, it simply is. But rather the moral and ethical applications, of the creation process.
hes talking about the multiple pro AI posts in this thread https://www.writingforums.org/threads/thousands-of-authors-urge-ai-companies-to-stop-using-work-without-permission.175706/page-2#post-2010353
I am pro-AI 100%, but I also intrinsically understand that it has limitations and the only way these limitations are going to be overcome is if we allow the experts to continue to train the models. To me this isn’t even about intellectual property. To me, this is about the many years from now when AI will have the capability to cure disease or solve complex issues that are beyond human scale that have real body counts attached. We aren’t going to be solving these issues, but if allowed, AI will and it will save real lives someday. Which is worth more to me than the prices of my current backlist and frontlist.
No one is arguing against training these systems. The argument is about how that training is being done. Ignoring copyrights, is a serious ethical violation, which is what upsets people and sparked the lawsuit. Yes the potential is there to solve problems that will save lives. That is not being debated.
I'd agree that AI in the right place has the potential to do great things but there is no connection between training one to write novels, and the involvement of AI in life saving bio tech etc Personally i suspect that AI bots are a red herring in that regard...I'd say the life saving tech is much more likely to come from breakthrough in tech not yet made which will make AIbots look like an abacus
AI can now pass the bar exam. It is also catching thing human doctors have missed. Which demonstrates that it is a useful tool. Like any tool, if used improperly it can cause more problems than it solves. The real question to me is can the dumb monkeys figure out how to use it properly.
on the other hand we have this kind of thing https://www.forbes.com/sites/mollybohannon/2023/06/08/lawyer-used-chatgpt-in-court-and-cited-fake-cases-a-judge-is-considering-sanctions/ Its problematic when an AI can't tell fact from fiction and invents cases when writing legal papers, so it might be able toi pass the bar exam but its someway from being able to argue a court case
Agreed, this seems to be a case of bias in the training of the AI. We monkeys are banging our thumbs with the new tool.
You know, AI got me thinking about something. The whole concept of technology evolving, becoming self-aware, and massacring the human race is purely fictional... something writers came up with to sell stories. And now that we're seeing finally seeing AI in real life, it's the writers that are fighting it on the frontlines! I suppose since we created the AI doomsday scenario, we're the most qualified to stop it. May we fight well!
At my new workplace we have monthly seminars where we can listen to a talk about a certain subject, be that a work-related application or something else. This month, by team manager explained to us the many applications of ChatGPT, but also the dangers of it and how to use it to our advantage. At first, I was very sceptical and slightly concerned about its abilities, but as the presentation continued it became clear, that the free AI chat tools, such as ChatGPT 3.5 and lower, are very limited in their source data. They have data from up to 2018, or something like that. Also, a valid point was made: What goes in, must come out. AI will be fed a few real books to begin with, which is wrong on all levels, and then new laws will be passed limiting what can be entered into AI, subsequently reducing its source material. In the end, it with trawl through the internet and pick up lots of AI-written books, thence reducing the quality of its output even further. I doubt that AI could kill off real writers even if it wanted to. In the end, it just copies what it already knows and jumbles it to suite the prompt. It’s very much like a photocopier; with each copy the quality worsens. So, even though I am absolutely against its use as a writing tool, it is fun to play with and ask it questions, but I am not afraid of it. It will never be smarter than genuine writers with good ideas.
Oh yeah, definitely. When my workplace presented us with their version of the tool, I asked it some very absurd questions, and got these excellent serious responses. I'll have to go dig one of them up real quick... There it is! This was from back in April. Haven't used it since. A few of these experiments had me fucking howling, let me tell you. The fact that the suggestions are already ~90% aligned with actual management emails was the cherry on top. But I still hate ChatGPT, and I loathe the idea of company emails being generated by AI. As if we don't skip reading so many of them as it is.
Now that made me laugh! That's why I find it fun to use, because you can literally ask it anything and confront it with the most absurd situations. But other than that, I wouldn't use it for anything else, and I am strongly against it being used to create work from other people's works, especially if they are illegally harvested.
Consider Adobe Firefly. Adobe is gaining a competitive edge in precisely this regard by training their AI on content they own. Images in their case. It's unlikely that there’ll ever be laws preventing big business from acquiring rights to various cultural produce and using it to feed AI. While we're discussing writers using AI for their novels, perhaps the more likely future scenario is non-writers using AI to create novels, films, songs, and more. Imagine this: instead of buying a pre-made book or movie, you come home from work and just click a few buttons. With a little guidance, an AI creates the perfect book, film, song or video game tailored just for you.
