AI generated, or workshopped?

[ShyMaryEllen]

A friend told me about how ChatGPT can critique writing in progress, so I tried it and she's right. It has given me some great advice to improve tricky passages, much in the same way as taking it to a workshop would do. Not adding things in, or writing text, but commenting on things like speech tags or punctuation.

My question is - would that count as work being AI generated when it comes to submitting to competitions - or anywhere, really? Is it cheating? I know that many, if not most authors take works in progress to workshops, and some submit pieces that have been marked for things like MA courses etc, so is this any different? I'm not sure what to think.

Seems like you are just using it to check your punctuation, so can't seen an issue with that really.

Thanks. It can go further than that, but that's all I've asked it do do with my work. I've put in cut and pasted extracts from other sources, and it does a full critique. I don't want to do that with my work in case it disqualifies it, but I suppose I'm asking how that differs from taking it to a workshop or similar.

AI is a huge issue in the writing community and publishing industry right now and my advice as someone in it is do not use it at all for any reason. There are numerous reasons for this, but the main one is that it will not make your writing better and will likely make it worse as it atrophies the creative muscles that genuine workshopping would strengthen. It is a slippery slope, and if you're allowing it to change your speech tags then you are AI generating your work in part.

LLMs have been developed by stealing the work of authors. It can create nothing original and it cannot help you build your own authentic voice. People in the industry feel very strongly about AI and the deluge of slop currently flooding agent inboxes and self publishing. No one reputable will touch your work if you have fed it through chatgpt so please, for the sake of your own creative process and chances of success, do not go near it.

It is a million miles away from workshopping or getting an editor. It's based on theft, it is extremely limited, it sucks away everything that makes art worthwhile and resonant. Writing is built on human experience and human connections. Keep going and you will end up somewhere far better with yours than AI could ever take you.

I know that (I have an MA in CW) and wouldn't want to have AI create content for me, as it defeats the object, apart from anything else. What I am asking really, is the difference between a critique from a computer and one from a group of peers. Suggestions are made in both cases, which makes the work collaborative if they are followed, and it is the finished result that publishers are looking at.

I am uncomfortable about it, hence the question, but have been musing over the moral aspect, rather than the professional one. If asking peers to critique something is expected, not just acceptable, why is asking Chat GPT to do the same wrong? I completely agree about getting it to generate writing, but that's not what I'm talking about.

Try it? I don't want to post anything that identifies my own work, but I've just run an old poem through it and this is an extract of the feedback:

I admire the ambition of this break more than I entirely admire the effect. It does create suspension and emotional emphasis, but visually the isolated dash beginning a line risks feeling mannered — the only moment in the poem where I become strongly aware of the constraint itself.

Someone in a workshop group could have said that, so I am not sure that there is a difference between listening to them and to AI - I would make my own artistic decisions anyway. All the same, I'm aware that there are strong feelings about it, although I'm not sure about no-one reputable touching anything that's been near AI. There is no reliable way to tell if it has advised, rather than written something, for a start.

But chatgpt is not critiquing your work. It's regurgitating a string of words that sound appropriate, and maybe even impressive. It has no understanding or appreciation of what you've written. A person offering a critique would be giving you an actual response to your writing. Chatgpt can't do that. Its advice to you is useless. It isn't actually 'aware' of anything, it's mimicking how a person might respond to you when it says that. It can't analyse your work and it absolutely cannot offer you anything truly helpful. It's an illusion.

And I promise you, every author and agent and editor I know (and it's a lot!) has the lowest opinion possible of AI when it comes to the writing process, and I cannot warn writers enough not to use it in any capacity. It's hard enough to break through, any hint of AI involvement anywhere in your work just gives agents/publishers a reason to reject you.

That makes sense. I've only recently started using AI for anything, and am new-fangled with it - I still thank it for its help🙄 - so I suppose I forget that it's not 'real'. You're right though. I think the best it can do is make you consider things you may not have done, in the way tarot cards might.

I have marked many essays/dissertations that I've suspected had had more than a little input from it though, and the sad reality is that even with the sorts of tools that universities have at their disposal it's remarkably difficult to prove.

The tarot card analogy is a good one - but tarot cards aren't destroying the planet with giant data centres that drain entire communities' water supplies. So I'd definitely stick to tarot!

No, there is no reliable way of proving it. But if you get representation, or a publishing deal, you'll sign a contract that asserts you haven't used AI in your work (look up the Shy Girl scandal to see what the impact is if AI use is revealed further down the line! It costs everyone something - the publishers take a financial hit, the author's career is ruined). Think about how it would feel to sign that contract with any doubt in your mind as to whether it's true. Think about how you would feel putting a 'human-authored' sticker on your book or a statement at the start asserting no AI has been used in the writing process. These things will become standard, and of course some people will lie. But they'll be taking a risk - and for me personally, it would feel way too icky. Writing books is wonderful, and it's better to do it in a way that doesn't make you feel at all compromised. If you'd feel odd admitting it to a future agent, editor or reader - don't do it.