Artificial Intelligence > Likes and Comments

Is this a good place to ask readers about AI and how they feel about it in the context of works of published literature from modern authors? Not necessarily the writing itself, but all that goes behind and into writing and getting published etc. I'm interested to hear what people think...

You can ask - just mind the advertising rules barring authors from talking about & advertising their own work.

Personally, I think AI has no place being in a book that is available for publication and consumer purchase. If an author isn't skilled enough to write their own story, there are other options available - a team of editors to clean up the mess, or a ghostwriter, or a writing class, or maybe just a different career altogether. But to me, using AI to generate a book (I won't say write, because it's not writing), is lazy, manipulative, and wrong.

Plus, I'm a critical reader - if your AI generated book slips through and I happen to read it, it will not go well in my review, even if I don't recognize it as AI at the time. I do not believe in the "if you have nothing nice to say, say nothing" review style. If I read your book, it's because it's available for purchase, and I will review it as a product.

I've read a few books over the last couple of years that, in hindsight, I believe were generated or at best "revised" with AI, and they were TERRIBLE. One of those was an independent author, and one of them was a traditionally published and very well known and highly respected author. Both got 1 star ratings, and LOOOOOOOOONG ranting reviews. And, again, I didn't even suspect that they were AI as I was writing the review, only afterward.

Not only that, but I have a friend on GR who seeks out and reviews AI generated books, and calls them out for the slop they are. He is able to immediately tell most of the time. There are ways to determine when a book is AI, and readers are getting more and more savvy about it.

In other words, I don't recommend it.

I am fully behind you on your point about the written word. Actually the same applies to AI generated movies, scripts etc. It should originate from human creativity and the magic that comes from spontaneous trains of thought.

You don't comment about the other things. Book cover art, research and targeted marketing to name just a few things. As an example, from what I understand from authors who use KDP to self-publish, you can choose a book cover from stock images, but these don't always fully resonate with the story. Not everyone can commission an artist to do a cover, and so potentially AI is a liberating tool. You might also argue the same for book trailers.

I posted a question a while back about audiobooks and whether or not it's a format making inroads into people's busy lives. Again not a cheap options for authors looking to reach readers and to bring the written word to life for some, and potentially - though I have not come across it yet - something AI would be good at doing. I have an audiobook myself but for personal reasons I chose to use a real actor. It worked well for me but again, if you don't have the resources "maybe" it's an option...?

This might sound rude, but, I personally don't care how expensive it is to produce a quality product - that includes the cover art and the audiobook (if producing one). While I completely understand and empathize with the idea of AI expanding access to aspiring authors, I fundamentally reject the idea that it produces anything worth real value that I would want to purchase as a consumer. It steals the intellectual work and actual labor of those writers and artists who came before, and parrots back a bastardized version. Furthermore, it takes paid work away from creators and artists, and the result is an inferior product. You get what you pay for.

I stopped reading self-published books a long time ago because I found that adage to be true. Without the skilled labor of editors and artists and whatnot, yes, the books were both cheaper to produce, and cheaper to buy... but they also were not good, and time after time I would be disappointed. It's not worth the effort of reading something that I know isn't going to be properly edited prior to release, or reviewing it when it is entirely likely to change based on reviewer feedback since they can just go in, update, and republish/repush a new version. (This happened to me several times.)

I will never willingly listen to an AI audiobook. Never. A skilled audiobook reader is an amazing asset to the story. There's no way that an AI system could compare.

If I see what looks like an AI cover, that's a red flag for me. If you use AI to generate the cover, how am I to know you didn't use AI to generate the whole book?

I also disagree with using AI for research, given the fact that the "analysis" AI generates is questionable, and I don't know what kind of biases or constraints or logic has been programmed in, or what data has been used to "train" it. Maybe the info is right... but unless you do your own research to check it, how would you know?

As far as targeted marketing, I am fine with that as an opt-in on the consumer's part. If they are fine with their data and habits being tracked to provide more specific marketing to their interests, then sure. But it should be opt-in.

I guess you have a very hard line on your view, and that's fine. It's not rude, it's just your clear opinion which I respect. We all have opinions and that's why I put the question out there... I do hope others will chime in and express views too.

For myself, I think it's more nuanced and I am loathed to cross off all self-published authors because of the risk that they might have used AI somewhere. I know two writers who have used AI but whose writing and edited books are 100% there own. One used it for a trailer, the other for a cover. They write because they love to tell stories. Should only those who can afford it, or are lucky enough to have a publisher, get to stand out and be seen and read? I am not so sure.

As for stealing IP - I get that. It's a mine field!

Re marketing. As you know, many of us who use the popular search engines have signed up to have our data / browsing behaviour used in targeted ways, whether we realise it or not. That's how they make their money. Furthermore, ads now adapt themselves in real-time to how we respond to them. That's AI too and you would never know. Is it wrong as an author to use / sign-up for that if you can reach genuine audiences who search for the genre you have written? Though I did hear somewhere that there are search engines now which promote the fact you remain anonymous, which is interesting...

Well, here's an aspect we haven't touched on yet. This entire conversation feels predicated around the publisher, not outrightedly admitting that the work was produced by AI. So we have to assume the reader does not actually know if it was. So maybe what this really comes down to is that if it's so easy to flag AI writing then it would just be grouped in with bad writing, assuming it is bad writing. So with that assumption, it's also a derivative that AI writing would fail in the mass marketplace. The original question relates to asking readers how they feel about AI, but maybe we need to ask how they would know if it was or if we're making assumptions that we can tell from reading it.

I think this is where I come down a little differently. I understand the accessibility argument, and I don’t think every use of AI is necessarily the same. I can see a distinction between using it for something like brainstorming, research assistance that is then properly checked, or even some of the more peripheral aspects of publishing. Marketing is a whole other conversation, too, because AI is already woven into so much of what we encounter online.

But I do think there’s a point where accessibility for one person comes at the expense of another person’s work. If AI allows an author to avoid paying an illustrator, editor, narrator, or other creative professional, then we’re not just talking about making publishing more accessible. We’re also potentially taking paid work away from the people who make their living doing those things. I don’t think the fact that someone can’t afford to hire them changes the value of their work.

The audiobook example is a good one for me. I understand why an author who can’t afford a narrator might see AI as a way to make an audiobook available. But a narrator isn’t just a voice reading words off a page. They’re interpreting the work. Their voice, timing, inflection and choices become part of the experience of the book. I listen to a lot of audiobooks, and that human interpretation matters to me.

And for me, the biggest issue is still the actual creative work. If I’m reading a book, I want to see the author’s mind and soul in it. I want the strange little choices, the individual voice, the perspective that could only have come from that particular person. That’s what makes literature interesting to me. That’s what makes one author’s work different from another’s.

Technical quality isn’t what I’m buying when I buy literature. I can understand wanting to tell a story and needing help getting it into the world. I can understand not having the money to hire an artist or narrator. I can even understand why someone might see AI as a useful tool in some parts of that process. But when AI starts replacing the human creative contribution rather than supporting it, that’s where I have a real problem with it.

And I don’t think it’s necessarily contradictory to feel that way while acknowledging that AI is already being used in search, advertising, recommendations, and all sorts of things we encounter every day. There’s a difference, to me, between an algorithm helping decide which advertisement I see and a machine generating the creative work I’m being asked to buy as a reader.

I suppose that’s ultimately what bothers me most. I don’t want a technically competent product. I want to encounter another human being through their work. I want to read something and feel that person’s imagination, experience, voice and personality on the page. If AI has taken the place of that person in creating the work, then for me, something important has been lost.

Thanks Vincent for adding to the conversation. Though my question was / is broader than just the writing part, I can fully relate as to why AI in the writing element might be the key thing for readers. That makes a lot of sense. After all, don't we want to read something that somebody took the time to imagine and visualise and then articulate in their unique way?

So it's an interesting point you make about whether an extensively AI written work would succeed in the marketplace and how many of us could tell. As of right now I think it might already be quite difficult for many people to spot some of the "better" ones. Here I mean the AI engines at the cutting edge. Things are moving exponentially so if it's already hard to tell, without properly enabled regulation and obligatory signing it will soon be almost impossible. Give it five years or less?

Will this mean that only well known authors doing all their own work will rule because they are trusted by readers, and the rest of us will fall by the wayside? I hope not! Possibly the onus will shift onto the publishers and platforms distributing the content as we start to see already. When I was submitting my work to agents three year's ago they ALL stipulated that the work had to be 100% original with no AI involved at all (amongst a host of other requirements I might add!!). If it will be left to publishers and distribution platforms, the irony is that before long, they too will rely on AI for that.

Good points about whether AI takes work away from people currently paid to do certain things. Art, narration etc. Undoubtedly in many cases it does. On the other hand, if resources is the issue, these same people would never spend the money and the work is not there to be lost. As it was mentioned earlier - it's a different question if AI is appropriating IP that others can exploit for free with no recognition to the originator. Not sure how we solve that one and we see cases in court about this all the time.

Loving this discussion!

I have just remembered that I received a 1-star rating on Goodreads from someone whose only comment was "AI cover". I do not mind low scores on the 5-star scale (thankfully I only have the one so far) as it's to be expected that not all stories appeal to everyone. However, not only is the reader's comment wrong - I never use AI in my books - it's very unhelpful.

I mention this because it comes back to the point made earlier in this conversation about "how you can tell when AI has been used?" Fair enough, raise concerns if you think it might be the case. especially when no mention of it has been made and you think deception is at play, but unless it's cast iron, NEVER accuse an author of that in their work.

Does anyone have a list of useful tell-tale clues to share? Could be for writing, covers or audiobooks. Would love to know. Here are mine:

Writing: over-use / string of fancy words that do not add anything / have no real purpose

Covers: I honestly do not have any. It is extremely difficult for me to tell. Video is possibly slightly easier because often there are elements that are misplaced or actions that don't work

Audiobook: monotone with low variations in pace, emphasis and pauses that give natural speech it's power

Becky wrote: "This might sound rude, but, I personally don't care how expensive it is to produce a quality product - that includes the cover art and the audiobook (if producing one). While I completely understan..."
Do you use spellchecker and a grammar checker? Do you use a search engine to do research? Then you're using AI.
Let me ask you a question:
Is it more ethical for an author to use AI or to use a ghostwriter?
After all, he's putting his name on something he didn't write.

Will wrote: "Do you use spellchecker and a grammar checker? Do you use a search engine to do research? Then you're using AI.
Let me ask you a question:
Is it more ethical for an author to use AI or to use a ghostwriter?
After all, he's putting his name on something he didn't write."

This is a wild statement. It implies that AI is ubiquitous and inevitable and unavoidable, and that is not the case. While it's true that many, if not most, search engines have incorporated AI summaries into their results, it's still possible to exclude them from the search (using -AI at the end of the query/search terms), and there are search engines that do not use AI at all.
www(dot)zdnet(dot)com/article/google-search-alternatives-no-ai/

Regarding your analogy to ghostwriters... Not only is this incorrect, it's disingenuous. A person HIRING, and therefore PAYING, a ghostwriter is not behaving unethically. The actual labor is being compensated and credited, even if that is not apparent to the consumer on the cover of the book. It is usually included in acknowledgements or on the copywrite page, but more importantly, it's in the contract guaranteeing them payment for their labor and skill.

An author generating a book via AI, which has scoured the internet and centuries of literature and written works to steal the effort and work and sometimes lifetimes of study and experience of humanity is not at all the same as hiring someone, and compensating them for their labor and abilities. This is an appalling comparison.

Julie wrote: "I don’t want a technically competent product. I want to encounter another human being through their work. I want to read something and feel that person’s imagination, experience, voice and personality on the page. If AI has taken the place of that person in creating the work, then for me, something important has been lost."

THIS.

Will wrote: "Is it more ethical for an author to use AI or to use a ghostwriter?"

IMO it is more ethical to use a ghostwriter because in that circumstance, editorial decisions are still being made by a human being. I do think it is unethical not to give credit to a ghostwriter.

If AI is, for all intents and purposes, the author of piece of prose, that fact ought to be identified on the cover of the book, so the buyer is forewarned.

Right now AI might not write all that well, but given time, I expect it will get to be quite good. Look at what consumer AI can produce in the field of pop music. Some of it's A LOT better than the garbage being put out by most of the humans.

If AI can produce what consumers are willing to pay for, human artists ought to sit up and take notice.

Here's another thought... and maybe in the end it does come down to what is lost. Like an actor's ability to say the same lines in a half dozen different ways based on where they choose to put the emphasis for an effect they are looking to get.

As humans, over a lifetime we learn from others, their experiences and countless observations we make along the way. When we read we are influenced. When we look at art or listen to music we are influenced. If we were then to become writers, artists or musicians, we might then opt to produce work that feels familiar whilst not replicating it, based on what we liked or admired. We do want to be unique but we cannot help but be shaped by our experiences and the things we respond to.

If I were an AI I might say that my scouring the internet to understand a topic and then to use it to produce something unique is just an accelerated version of the same thing. As an AI, thousands of datapoints are used to "do something" unless someone asks it to "do it in the style of", and even then I think AI would say they are not "copying" someone, rather they are looking at characteristics of their work.

Do we judge one to be okay because it's natural human behaviour and the other wrong because it's contrived?

I did an experiment yesterday with ChatGPT and I can say that the painting it produced, whilst good, was nothing like the artist's work that I asked it to "copy". At the same time - maybe because I know how it was done - whilst it was good it also felt stilted. Lacking a soul in some way as if it were too perfect.

And I think think this is the essence of the point. A current weakness of AI. In looking at too wide a set of data, or at technical characteristics of a piece of work, it cannot understand what humans find an emotional connection with. It tries to learn from everything, and as a result produces something that reads, looks or sounds "good" but misses some kind of harmony that resonates with the human mind - and misses. Trying to please everyone but in the end not being exceptional for anyone.

But as I mentioned earlier, give AI another few years and sadly I think even this element will become almost invisible, AI tuned to us as individuals.

So it will all come down to this... as Julie put it. "I want to encounter another human being through their work."

Some of us will feel that way and others will not care so long as they enjoy the story, the image, the sound.

Luigi wrote: "Here's another thought... and maybe in the end it does come down to what is lost. Like an actor's ability to say the same lines in a half dozen different ways based on where they choose to put the emphasis for an effect they are looking to get."

Sure, and that would definitely strongly impact audiobook performance, but what else would be lost is the author's human ability to convey written meaning and context and subtext and nuance, to convey subtlety and leave aspects open to interpretation and the READER's experience and understanding. What would also be lost is an author's ability to grow and change and shape their work through that new experience. I just don't see how it would be possible to do that via a chat prompt, no matter how detailed it is. (And one could argue, if it's detailed enough to convey the meaning and output you intend... it's detailed enough to just write it yourself.)

I can't tell you how many times I have re-read a book at different stages of my life and felt drastically differently about it, because the way I see the world has changed and evolved. Those situations are usually due to aspects of the author slipping through, intentionally or not, and impacting my experience with their work. Maybe they left some elements or aspects vague and open to interpretation and my interpretation changed. Maybe they subconsciously wrote in veiled racism or sexism or xenophobia, or used stereotypes, and didn't even realize or intend that to come through. Maybe they carefully selected every single word, and were meticulous in their language and intention. Those choices, and subconscious human elements, will be lost, too.

100% agree that like some very good scripts / movies too, the best story telling leaves plenty of room for people to take different things away from it, and to discover new things on multiple readings / viewings. I've lost count of the number of times I have read TLOTR and found things I don't recall noticing before.

It's also true that writers leave a lot for the reader's imagination to fill in as well. Yes it's the writer's story to tell but it's also the reader's journey to discover and see for themselves.

Like someone else said earlier in this chain of conversation, AI will get there eventually. We might not see or understand how, but then who would have thought just 10 years ago we would be where we are now. So long as content is properly labelled I think there is space in this world for both. What we should fear is the hidden hand.