the use of AI to sculpt your writing

[Genly]

TBH, I am not sure when I have ever had to research something so in depth, that I found Wikipedia not to be helpful.

AI has led me on some wild goose chases with IT, but not many.

One issue I had with AI, was a long goose chase, that when it ended, I wanted to be helpful and say, Hey, AI, the actual answer was thus....but it would allow for that. So, my solution will not likely be useful to others.

In case the internet is watching... Yes, you can get an Office 365 machine, to connect to an old local Exchange 2016 server by disabling the internet gateway. And, if you alter the hosts files to point to the local machine, it will not screw up again after internet is restored. Only spent a day of screwing around with registry keys to get to that. But...its pretty apparent, MS wants people to be on the cloud.

As someone who does a lot of research for writing both fiction and non-fiction, I find Wikipedia to be helpful, but only as a starting point. I still have to access the primary sources, just to check that Wikipedia is correct -- because sometimes it isn't.

The same goes for AI searches. I've had direct experience of noticing that an AI summary of a particular topic was quite wrong and out of date. I don't use them in my work for this reason.

As for using AI to write text: this kinda takes all of the fun and effort out of the process. It would be like letting a robot do your workout for you. Of course, at this time I'm not trying to make a living out of being a writer. Someone who is attempting to do so might have a much more negative opinion of AI.

 

[Quarney]

As a non-native speaker too, I think that depends on how you use ai--AI as a replacement, or as a collaborative partner? I mean, seriously, that I assume many people have prejudice about ai in which the only way to use ai is by "GPT, generate me some articles"--is obviously hilarious. In many academic researches or some creators might would choose to use ai in this way: some ideas suddenly hit-they organize the thoughts to ai-ai returns ffeedback for them-repetition-and then they check the final to ensure quality. I think this is the true human labor, machine is just a tool.
I usually write in English, throwing it to ai for feedbacks, or have a discussion about some topics/imposing questions to it.
And also, AI hallucination is not a moral failure--it is a structural flaw that requires users to overcome. I usually notice that ai would overcomplicate and give me out-of-date data, which require you to double check--that's also the pure human labor.

 

[Genly]

Maybe... but what if a reader of an AI response lacks the expertise to check whether it is accurate or not? At its current stage of development, AI can be seriously misleading. And "checking" is sometimes not enough.

As someone who actually does academic research regularly, I'd say that AI is useful as a sounding board. But that's about it. Probably it will get better, though, with time.

I don't know how useful it is for improving text. But if it works for you as a non-native speaker of English, then that's ok, I guess -- provided that all it is doing is correcting your grammar or maybe giving you some general ideas for content. If it is also correcting your style and imagery, then that is a problem.

 

[LanternHolder]

I use AI in a professional capacity all the time. It composes emails and technical briefs that previously would have been the bane of my existence. But for creative writing? As close as I'll go is using AI in writing exercises where it will generate a prompt that I'll use for a word sprint - just to get the creative juices flowing. Would I use AI to do the writing for me though? Where's the fun in that?

 

[Quarney]

Maybe... but what if a reader of an AI response lacks the expertise to check whether it is accurate or not? At its current stage of development, AI can be seriously misleading. And "checking" is sometimes not enough.

As someone who actually does academic research regularly, I'd say that AI is useful as a sounding board. But that's about it. Probably it will get better, though, with time.

I don't know how useful it is for improving text. But if it works for you as a non-native speaker of English, then that's ok, I guess -- provided that all it is doing is correcting your grammar or maybe giving you some general ideas for content. If it is also correcting your style and imagery, then that is a problem.

oh, I respect your worry. But as far as I know, many issues have made more nuanced regulations about ai using and disclosure. As for the accuracy, many issues say that author must be responsible for the content--that is to say, author have the responsibility to double-check the ai contents. Many authors have done in that way.
And yeah, I also use ai as a sounding board. Most of time, ai could just simply repeat your arguments(re-organize, polish, or impose some questions--based on the probability). But when it comes to style and imagery--well that's a grey world. But from my point of view, I think the source of your inspiration is totally neutral--no matter you r inspired by ai or a human. In this way, you can hardly tell whether if u r corrected by a human after adopt their opinion on your style or imagery. Well, it is quite common that some editors would ask you to change here and there, right?

 

[Genly]

In this way, you can hardly tell whether if u r corrected by a human after adopt their opinion on your style or imagery. Well, it is quite common that some editors would ask you to change here and there, right?

In my limited experience of preparing text for publication of fiction, editors do not correct style or imagery. They might correct typos or supply a better turn of phrase, or sometimes ask for a portion of the text to be deleted or even rewritten. Your style is your own, and if you get an AI to do it for you, then it is no longer your piece of writing.

Really, you should stick to using AI to fix any grammar problems in the text. For a non-native speaker of English, that would be a useful learning experience. Or, as others have said on this forum, you could try writing in your native language instead.

 

[Prince of Spires]

In my limited experience of preparing text for publication of fiction, editors do not correct style or imagery.

This very much depends on the type of edit and the quality of the editor.

Some editors focus solely on the big picture and the broad strokes of the manuscripts. Others focus on the specific way you say something. And others still focus only on the typos and grammar. And then everything in between as well.

And some editors have stronger opinions than others about how something should be said. They might try to impose more of their style on the manuscript or of what they feel like the style of the story should be.

Yes, a good editor will try to make the author write the best story he can in the style of the author. However, not all editors are good editors...

 

[Mad Swede]

oh, I respect your worry. But as far as I know, many issues have made more nuanced regulations about ai using and disclosure. As for the accuracy, many issues say that author must be responsible for the content--that is to say, author have the responsibility to double-check the ai contents. Many authors have done in that way.
And yeah, I also use ai as a sounding board. Most of time, ai could just simply repeat your arguments(re-organize, polish, or impose some questions--based on the probability). But when it comes to style and imagery--well that's a grey world. But from my point of view, I think the source of your inspiration is totally neutral--no matter you r inspired by ai or a human. In this way, you can hardly tell whether if u r corrected by a human after adopt their opinion on your style or imagery.

Writing this as someone who peer reviews academic papers I can tell you that we will scrutinise in detail any paper where we know or suspect that AI has been used. This is because we must be certain that the authors understand and can justify what they have written and the conclusions they have drawn.

As a researcher with a research degree I would never use AI in any of my work. It's my job to do the research and to document it in a way which reflects what I have done and which can guide others doing similar work.

Well, it is quite common that some editors would ask you to change here and there, right?

This depends on what sort of editing you are talking about.

When writing for an academic journal there is usually a house style, and once the peer review process is complete an editor will ensure that the paper concerned follows the correct style.

When your work is being edited for publishing then the editors will not usually change your style or your imagery simply because that is part of your prose.

 

[Quarney]

In this way, you can hardly tell whether if u r corrected by a human after adopt their opinion on your style or imagery. Well, it is quite common that some editors would ask you to change here and there, right?

In my limited experience of preparing text for publication of fiction, editors do not correct style or imagery. They might correct typos or supply a better turn of phrase, or sometimes ask for a portion of the text to be deleted or even rewritten. Your style is your own, and if you get an AI to do it for you, then it is no longer your piece of writing.

Really, you should stick to using AI to fix any grammar problems in the text. For a non-native speaker of English, that would be a useful learning experience. Or, as others have said on this forum, you could try writing in your native language instead.

It seems that we have the different definition of authorship and how style comes. What I said is that AI can sometimes be a source of inspiration, or a reference for your style. The ultimate decision is made by the author. The process is quite the same as a designer finding some references for their works.
Imagery and style can be influenced by so many factors, but you can develop your own numbnail of them. In this way, AI is the material--but you should have your own philosophical frames of what you should have and what don't.

 

[Quarney]

Writing this as someone who peer reviews academic papers I can tell you that we will scrutinise in detail any paper where we know or suspect that AI has been used. This is because we must be certain that the authors understand and can justify what they have written and the conclusions they have drawn.

As a researcher with a research degree I would never use AI in any of my work. It's my job to do the research and to document it in a way which reflects what I have done and which can guide others doing similar work.

This depends on what sort of editing you are talking about.

When writing for an academic journal there is usually a house style, and once the peer review process is complete an editor will ensure that the paper concerned follows the correct style.

When your work is being edited for publishing then the editors will not usually change your style or your imagery simply because that is part of your prose.

As I have mentioned, I noticed that many issues have developed more nuanced ai regulations--which is alighed with your ethical caring. And I respect your workflow, but I might also gently confirm that using ai is not a "false", but another choice for studying and researching. AI using should be regulated--but we are not regulating "AI", but how Humans use AI.
And yeah, I think my example of editors is quite...oversimplified. I would like to change to another one: would we call a writer as "cheating" simply because he chose to simulate a specific style from another "human" author? In this way, the "another" one has influenced, or even "correct" their style. But the Premise is always that the writer themselves decided to be influenced or corrected. The process is quite similar when someone finds inspiration in ai.

 

[Karlin]

It seems that we have the different definition of authorship and how style comes. What I said is that AI can sometimes be a source of inspiration

Inspiration is a major part of my creative process. Why would I want AI to inspire me?

 

[Mad Swede]

As I have mentioned, I noticed that many issues have developed more nuanced ai regulations--which is alighed with your ethical caring. And I respect your workflow, but I might also gently confirm that using ai is not a "false", but another choice for studying and researching. AI using should be regulated--but we are not regulating "AI", but how Humans use AI.

What some (many?) call AI is in fact what is known as a Large Language Model (LLM). LLMs basically work by using conditional probability analyses and statistical inference to go through a data set and produce a result, based on whatever prompt was given. The quality of the result is entirely dependent on the quality of the data set - if anything is missing from the data set then the results may not be reliable. As a user you have no way of checking the data set. For me as a researcher and as a peer reviewer that alone is reason enough to not use AI.

For study and research I need previous research and results, which are best found using a good academic index like JSTOR. The I start reading. Data I can get separately.

And yeah, I think my example of editors is quite...oversimplified. I would like to change to another one: would we call a writer as "cheating" simply because he chose to simulate a specific style from another "human" author? In this way, the "another" one has influenced, or even "correct" their style. But the Premise is always that the writer themselves decided to be influenced or corrected. The process is quite similar when someone finds inspiration in ai.

You may think that some authors copy other authors styles, but when you start reading their works you see that all authors add their own unique perspective. I often take the example of novels about private detectives. Some would say that Dashiell Hammet and Raymond Chandler both wrote hard boiled detective stories whoich are similar in style, but when you read their novels you realise that their prose styles and the focus of their stories are very different.