AI Generated Books

[Barrington Boots]

I got a book for Christmas that is so awkwardly written that it reads like it was either generated by AI, or has been fed through a translation programme that's made it come out all weird.

Not naming names and edited slightly in case I'm making a terrible accusation againts a legit author, but here's an extract from literally a random page in the book:

"Behind the second army the <snipped> forces arrived. Horses both tamed and wild aided their speed. Like the many animals to follow us, they drew to their plight. Knowing forests, they swifted the warriors beyond path and road, eliminating countless miles. They galloped forth with a frightful torrent."

The whole book is like this - every single paragraph doesn't quite make sense. Dialogue is awkward, sentances poorly built (and there's tons of grammatical errors)

It may be that the author simply has a style that I don't like but I'm suspicious I'm reading my first AI-generated book. Last I experimented using an AI programme to write a couple of dialogue scenes for me that I wasn't feeling that I could then edit into something decent, and it churned out stuff a bit like this.

Anyone else come across anything like this?

[IndigoPrime]

Most AI writes better than that (if still painfully generically), so it might be an early book written primarily by GPT and without the input of an editor.

[Barrington Boots]

Most AI writes better than that

Burn!

Chat GPT is exactly what I'm thinking as the author here, yes, as that's the only AI I've had experience of.

Ironically my original post contains spelling mistakes and missing words so perhaps the writer of this just wrote it really fast between work calls.

[Colin YNWA]

No really directly related but I somehow watched a YouTube channel about a bloke who created AI generated colouring book and sold them over Amazon as print on demand and had developed a second income. All be it at the time of the video I watched not much return for the work involved but that was improving.

Watching that video I scanned the recommendations below and my gosh there seemed to be a LOAD of over channels covering the same or similar ideas!

[Barrington Boots]

My wife is in an author chat group and there's a huge concern there over AI generated books: it sounds like its a growing problem. Someone just churns out loads of them and whacks them on Amazon. The book I have has no reviews on Amazon, Goodreads or anywhere else I can find (although the author has a website)

I'm not sure AI colouring books is of the same concern but presumably thats just the same deal but leeching off someone elses art. There's a data poisoning tool called Nightshade thats now been released and a few artists I know are very keen on it, but I think its too early to say if it will make a difference.

[Hawkmumbler]

To paraphrase the irascible yet genius Hayao Miyazaki: "AI art-"

[broodblik]

I do not want to read a book done by AI, for me it is a form of fraud when you put your name to a "story" which you actually has not written. Entertainment must stay in the domain of our mere mortals.

[The Legendary Shark]

The Distinction of Creativity: Why AI Cannot Create Literature

Introduction:
Artificial Intelligence (AI) has made remarkable strides in various fields, but the realm of literature remains elusive to its capabilities. While AI can mimic certain aspects of human creativity, there are fundamental qualities inherent in literature that cannot be replicated by machines. The essence of literary creation lies within the depths of human emotion, experience, and imagination, which AI lacks inherently. As we delve into the complexities of language, storytelling, and the human condition, it becomes evident why AI cannot truly create literature.

1. Emotional Depth and Human Experience:
Literature evokes profound emotions, allowing readers to connect with characters and experience nuanced human dilemmas. AI lacks the intrinsic quality of human emotions—its understanding of sentiment is limited to surface-level analysis rather than genuine comprehension. Writing authentic narratives that resonate with readers requires the depth of human emotions, drawing from personal encounters, vulnerabilities, and the complex tapestry of life experiences.

2. Symbolism and Subtext:
Literature often relies on the intricate placement of symbolism, metaphor, and subtext to convey deeper meanings. These literary devices are grounded in cultural, historical, and emotional associations that are deeply ingrained in human communication. AI, programmed to recognize patterns and derive meaning from explicit information, struggles to grasp the subtle nuances and symbolical significance that human authors can effortlessly incorporate into their works.

3. Creative Imagination:
Imagination is the backbone of literature, enabling writers to craft worlds, characters, and scenarios that captivate readers. The imaginative process involves synthesizing disparate elements and translating them into unique narratives. AI, despite its remarkable ability to analyze vast amounts of data, lacks the imaginative spark that stems from human ingenuity, intuition, and creativity. The sheer unpredictability and novelty of human imagination cannot be replicated by algorithms alone.

4. Authentic Voice and Style:
Each writer possesses a distinctive voice, style, and perspective that shape their literary works. Writers infuse their compositions with personal idiosyncrasies, reflecting their cultural background, life experiences, and individuality. AI algorithms, while capable of mimicking established writing styles, cannot authentically embody the complexities and nuances of human expression. The innate subjectivity and personal touch found within literature are the hallmarks of human artistry that AI cannot genuinely replicate.

Conclusion:
While AI continues to transform many aspects of our lives, it falls short in creating literature that encompasses the depth, emotion, imagination, symbolism, and unique perspectives that make it truly captivating. The essence of literature lies in the human condition, the intricate dance between language and emotion, and the intimate connection forged between writers and readers. The future may offer advancements in AI, but the distinctive qualities that define literature will continue to belong exclusively to the realm of human creators.


Written by deepai.org two minutes ago...

[The Legendary Shark]

Unleashing New Literary Horizons: How AI Can Create Literature

Introduction:
Artificial Intelligence (AI) has brought about groundbreaking advancements in various fields, and the domain of literature is no exception. While some argue that the essence of literary creation is uniquely human, AI's ability to analyze vast amounts of data, understand patterns, and generate coherent narratives offers a compelling case for its potential to create literature. In this article, we will explore how AI can transcend preconceived limitations, challenge traditional perspectives, and push the boundaries of literary expression.

1. Enhanced Creative Possibilities:
AI algorithms can process and learn from an extensive corpus of literature, enabling them to identify patterns, themes, and narrative structures that have resonated with readers throughout history. By leveraging this knowledge, AI can generate unique and innovative storylines that blend familiar elements with fresh perspectives. Through the synthesis of different literary styles and concepts, AI can introduce readers to entirely new forms and genres, expanding the boundaries of literary creativity.

2. Efficient Ideation and Collaboration:
The creative process often involves iterative brainstorming and collaboration. AI algorithms can rapidly generate a multitude of ideas, plot points, or character traits, acting as valuable assistants to human authors. AI algorithms can analyze a writer's existing work, suggest improvements, and provide alternative storylines or character arcs. The ability to collaborate with AI can accelerate the creative process, helping authors refine their ideas and craft richer narratives.

3. Language Mastery and Style Adaptability:
AI technologies, such as language models, have made impressive advancements in understanding and generating coherent written text. These AI systems can adapt their writing style to imitate the voice of a specific author or even create new styles altogether. By analyzing vast amounts of literature, AI can learn to mimic certain writing techniques, such as stream of consciousness or poetic prose. This adaptability opens up exciting possibilities for AI to generate compelling and diverse literature.

4. Experimentation and Exploration:
AI's lack of human biases and creative constraints allows it to delve into experimental storytelling techniques that challenge traditional literary norms. Through AI-generated literature, readers can explore surreal, abstract, and unconventional narratives that push the boundaries of storytelling. AI can also generate interactive narratives, where readers become active participants in shaping the plot, creating immersive and engaging literary experiences.

Conclusion:
While controversy surrounds the idea of AI creating literature, its potential to enhance and expand the literary landscape should not be ignored. By leveraging its data analysis prowess, collaboration abilities, language mastery, and a vast pool of literary knowledge, AI can contribute to the creation of exciting and innovative literature that complements human creativity. The synergy between AI and human authors has the potential to unlock uncharted literary territories, making literature a more dynamic and diverse field than ever before.


Written by deepai.org one minute ago...

[The Legendary Shark]

Personally, I find a.i. useful (if often frustrating) as a research and imagineering tool but I'd never rely on it for actual writing. I is perfectly capable of being an writist myself, thankings very muchly.

[Hawkmumbler]

Considering AI script programs can only function as they do by stealing incomprehensible swaths of copyrighted material (more often than not, from small writers who go without recognition, credit or payment for their labours even before they're ground up and fed into the dread maw of the content mill) I really don't see how anyone can accept AI as a 'tool' without inherently admitting they're fine with content theft and plagiarism.

[IndigoPrime]

I think it depends what you're using it for. I know coders using it, with LLMs having cribbed code openly available on the web. There, I suspect, attribution will eventually be the problem.

I also use LLMs in my own writing, but for rewording paragraphs and occasionally for basic outlining. They're no good (yet) for the meat of any creative copy. For research purposes, Perplexity is interesting in that it embeds references. So rather than getting a load of hallucinations (ChatGPT's favourite thing), you get responses you can check and potentially integrate.

I don't at this point see big benefits for eg fiction writers though. Even from a proofreading standpoint, these LLMs are woeful. They miss so much.

[Barrington Boots]

That's interesting IP. I've experimented with using GPT for coding. It was that that led me to use to to write a bit of dialogue - I was working from a similar state of mind that GPT could give me a framework that I'd then edit into what I actually wanted.

In both cases it was a bit useful to give me a kickstart on something I wasn't confident of, but not really a good substitute for doing it all myself and I eventually stopped. With this book, what I think someone did was write a plot summary into GPT or similar and got it to do the donkey work for them. I've absolutely no idea why anyone would do that tbh as it just produces some weirdly phrased crap like the paragraph I posted above, and surely you'd need to sell a LOT of print on demand books to make it worthwhile. Nobody who read one of these books is going to read another by the same author. It all feels a bit self-defeating.

With things like literature and art, my own view is that there's no place for AI in it. The spark of creativity isn't there. I'd never knowingly purchase anything using it, and I've stopped supporting stuff that has. That's even before you get into the morality of it: I know artists who rely on comission work and whose work is being scraped and stolen and it is very upsetting for them. That's another debate though.

[IndigoPrime]

One question that's rarely asked is what happens when the source material runs dry. I see arguments from publishers that LLMs should be used for more 'basic' writing, such as news reporting. But that relies on a source existing for the LLMs to draw from – and ideally multiple sources. Similarly, I see people claiming that soon we won't have search engines that return results for eg reviews. Instead, they will directly provide recommendations, based on amalgamating what exists out there. But if no-one's going to the original sites, which can no longer support themselves with ads or subscriptions, those sources simply won't exist.

As for art, I'm in much the same space as you. I don't at this point see any value. At best, LLMs can be used for 'shaping'. If you've already done your research, and have an outline, you can chuck that into an LLM and get something formatted back out, potentially cut back. But even then, it needs a skilled editor to work with the output and make it good. I've been writing professionally for quarter of a century now, and so am confident in my ability to do that. My concern is people are going into this green (or being forced to by management), and the results there are clearly not going to be sufficient. (And then you get the nadir of writing such as in the book you mention.)

[The Legendary Shark]

This idea of training your own a.i. is interesting.

Imagine if Pat Mills fed everything he'd ever written into his LLM, then all he'd have to do is put in an outline and sit back while the program filled in the blanks. And maybe Brian Bolland could do the same with his artwork, or Jim with his lettering. That's a huge increase in productivity and, of course, for a few extra quid I'm sure the above luminaries would provide entirely human-based work.

The technology needs to improve drastically, though, as I suspect it will.