AI written books

[gmw]

Originally Posted by Quoth

But it will be plagiarised from multiple human written sources. That's how the LLM works.

Certainly LLM are trained against existing sources of text - but so are humans. Even dictionaries and grammar rules are constructed by referencing existing sources of text.

In LLM the training establishes models and probabilities for how words and sentences hang together, and assuming your training material is large enough (the first "L" in LLM) the chances of it parroting any source training material directly is supposed to be very small.

If you have sources demonstrating that current systems are producing recognisable reproductions from their source training material then by all means share the links, it should make for interesting reading.


My understanding is that the accusation of plagiarism via AI is normally about asking an AI to restate some other text (another person's essay, for example) so that your (actually the AI's) copy of that text doesn't look like the original.

[ZodWallop]

Originally Posted by Quoth

But it will be plagiarised from multiple human written sources. That's how the LLM works.

I think you are making too big a deal out of this without understanding how learning works. Are Warhol's Mona Lisas or Roy Lichtenstein's comic paintings plagiarism? And those are purposely obvious examples. Every artist learns from what has come before.

I'm not wild about AI produced books. But you claiming it is plagiarism every ten posts isn't convincing anyone of anything.

[DiapDealer]

Originally Posted by jhowell

It could churn out an endless steam of books you love with characters you connect with.

And if it did, I honestly don't think I'd care all that much how it came into being.

(Ignoring for the time being the "connect with" thing. That's already been hashed in another thread )

[SleepyBob]

Originally Posted by DiapDealer

And if it did, I honestly don't think I'd care all that much how it came into being.

(Ignoring for the time being the "connect with" thing. That's already been hashed in another thread )

On an individual level, I can appreciate that. On a societal level, I think it would run the risk of causing stagnation, drowning out attempts at innovation by human authors.

A musicGPT AI that was only ever trained on classical music is unlikely to develop blues, ragtime, ska, or rap. And when AI can generate good enough material for pennies, no human author or composer would be able to make a full-time living from their craft.

[DiapDealer]

I tend to not worry too much about the sky falling because of short term paradigm glitches. Automation was supposed to put all laborers out of work by now, too. This too shall pass, and good human authors' works will get published for money.

Status quo interrupted by short periods of upheaval was, is, and ever shall be. *shrug*

[ekbell]

Originally Posted by gmw

First, what's the point of a publisher if they're not going to check over submissions? These days one of their self-proclaimed reasons for being is to act as a gate-keeper and filter.

Second, I don't think that a 'Does this plot make any sense' check helps you separate AI from human - seems to me I've come across plenty of (assumed) human texts with nonsense plots.

Third, take a look at the the earlier critique of the ChatGPT example and tell me you've never seen the same criticisms of human text. Indeed that post also critiqued another post too, on not dissimilar lines, and yet I give Quoth the benefit of the doubt as regards being human.

Panic may be inappropriate, but there is certainly enough going on in this space to make it interesting to watch.

My point was that it's hard enough to read through a pile of human written dreck to get to the one decent work, adding an pile of A.I. written dreck will lead to the slush pile growing faster then the slush readers can possibly read.

It matters not if the regurgitation of half digested material gathered from dubious sources is A.I. produced or lovingly crafted at home by human hands and mind, what matters is the sheer amount of the stuff.

[Quoth]

Originally Posted by DiapDealer

I tend to not worry too much about the sky falling because of short term paradigm glitches. Automation was supposed to put all laborers out of work by now, too. This too shall pass, and good human authors' works will get published for money.

Status quo interrupted by short periods of upheaval was, is, and ever shall be. *shrug*

That's mostly true.

The term "Luddite" often is used to describe someone who is opposed or resistant to new technologies. In reality they wanted a slice of the greater profits (x100 or more) generated by the mechanised and automated Mill owners. The workers in the Mills had very poor pay and working conditions.

An issue is the dishonesty of the people seeking VC investment and the dishonest claims of AI. It's getting worse.

https://www.theregister.com/2023/02/..._agi/?td=rt-3a

Open AI now making claims about AGI that are wrong on two levels.

Note: Anyone can search and read that anything I've posted on this subject is simply reporting. I'm not writing wikipedia articles or journal papers so I'm under no obligation to give sources. Nor am I prepared to post my CV here. Personal attacks on the messenger are uncalled for.

Nor do I agree with Charlie Stross:
http://www.antipope.org/charlie/blog...your-bets.html

Saw it 4 days ago here:
https://historians.social/@cstross@w...25845889562171

[j.p.s]

Originally Posted by gmw

I think this suggests that AIs could well produce books that at least some readers will find acceptable, not because they're great stories by literary standards, nor because they are producing original material, but actually because they are reproducing patterns we find pleasing.

Doesn't that describe a significant chunk of human written books, some even by major authors?

[Quoth]

Originally Posted by j.p.s

Doesn't that describe a significant chunk of human written books, some even by major authors?

All Mills & Boon

[ZodWallop]

Originally Posted by ekbell

My point was that it's hard enough to read through a pile of human written dreck to get to the one decent work, adding an pile of A.I. written dreck will lead to the slush pile growing faster then the slush readers can possibly read.

Don't worry. The A.I. editor-bot will sort it all out in a matter of seconds

[DNSB]

I was on a website where one person posted that they were using AI to "help" write children's books for Amazon. Their claim was that in the last 6 weeks, they have produced 23 books which have earned an average of $150 US so earnings were $3450 US for the last 6 weeks. The user name did not match any authors on Amazon so I couldn't look for examples.

[gmw]

Originally Posted by DiapDealer

I tend to not worry too much about the sky falling because of short term paradigm glitches. Automation was supposed to put all laborers out of work by now, too. This too shall pass, and good human authors' works will get published for money.

I agree with the part I bolded (the rest is best associated with my comment below). In this case human-to-human (or appropriately authenticated) contact will still see human produced material make it through. And I don't think it is necessarily a bad thing that at least some interactions get encouraged to take a more human-centred approach.

Originally Posted by DiapDealer

Status quo interrupted by short periods of upheaval was, is, and ever shall be. *shrug*

I know it's not your thing to get worked up about stuff, and that's cool man, we dig it. But no one is claiming this as the end of the world; a new status quo will be established in time. Some will gain by it and some will lose out, as "was, is and ever shall be." But it's still interesting (to some of us, anyway) to discuss what things might change and what things might get lost along the way.

[DiapDealer]

Originally Posted by gmw

I know it's not your thing to get worked up about stuff, and that's cool man, we dig it. But no one is claiming this as the end of the world; a new status quo will be established in time. Some will gain by it and some will lose out, as "was, is and ever shall be."

I don't know, man...

Originally Posted by SleepyBob

On an individual level, I can appreciate that. On a societal level, I think it would run the risk of causing stagnation, drowning out attempts at innovation by human authors.

A musicGPT AI that was only ever trained on classical music is unlikely to develop blues, ragtime, ska, or rap. And when AI can generate good enough material for pennies, no human author or composer would be able to make a full-time living from their craft.

The above sounded just a touch End of the World-ey to me. *shrug*

Originally Posted by gmw

But it's still interesting (to some of us, anyway) to discuss what things might change and what things might get lost along the way.

By all means, carry on my wayward peeps. My last was a direct reply to the cat who seemed to be stressin' over the notion of future societal stagnation because of this trend. Not a general suggestion that y'all should mellow and consider vacatin' the scene.

[Pierre Lawrence]

More on AI neural networks vs. the brain:

https://nautil.us/why-the-brain-is-so-noisy-237301/

[Quoth]

https://www.theregister.com/2023/03/...mful/?td=rt-3a