I can't do that, Dave: AI drowns top sci-fi mag with story submissions Science fiction and fantasy periodical Clarkesworld Magazine has temporarily paused submissions from authors after being inundated with AI-generated stories. Launched in 2006, the monthly Clarkesworld publishes a mixture of science fiction and fantasy short stories, articles, and interviews. The award-winning mag is known for …
Probably not all books. There are still people who ride horses, people who make pottery, people who knap flint. (I've done the latter two myself, in fact, and plan to learn to ride one of these days.)
But it's very likely that a great deal of language expression will be machine-generated in the near future, yes. That's no surprise. Icon Group International (Phillip Parker's publisher) has been selling computer-generated books since the turn of the century. Computer-generated prose has been used extensively in journalism and some other domains for some years now. Everyone who was paying attention knew this was coming.
Purely human-written prose and verse1 will likely become something of a luxury good.
That said, we've raised a couple of generations who are the most literately productive of any in human existence. Millennials and Zoomers have done far more writing than previous generations, and Gen X is just behind them. The vast majority of that writing is informal and either interpersonal (e.g. texting and message apps) or published for free online, but even so that's a habit which may prove hard to break. So even if professional writing is largely taken over by machines, it's possible that informal writing will still mostly be done by humans for some time to come.
1Based on the output of LLMs thus far, I think human-written poetry is safe for a while yet. However, most readers, even educated ones like Scott Aaronson, don't do a good job of distinguishing between adequate (much less good) poetry and mere verse.
With reddit style updoots, people vote on the shorts they like and the curators skim those with the highest score for review and publication? Crowdsource your review process.
I mean, unless someone was going to go to the lengths of creating many automated accounts to upvote their stories and downvote others to try and push their ai generated story to the top of the list to then be ignored by the reviewers..maybe some captcha v3 style monitoring would help with that.
No, not that. User moderation only works if it's consistent - ie the same users always check every submission. Like, you know, a team of editors ;)
If you don't have that then it just becomes the usual social media tit-for-tat "Likes" game, where the people who spend the most time liking other people's posts get the most reciprocal Like-as-a-thank-you. The people who succeed in that game are the ones who spend the most time clicking boxes rather than generating decent content.
If you are interested enough, you can get a set of friends who all like what they're told to like. This can work if you can't actually get a bot army to do it for you. When you come down to it, why have a chatbot write a story, anyway? There are only two reasons that make sense: you want to get paid for writing a story and you don't want to do the work or you want to show off that you wrote a story and you don't want to do the work. Both are subject to artificial voting.
In addition, there's not that much incentive for non-artificial voting. Would you like to read ten thousand sci-fi short stories, a few thousand of them generated by a chatbot, a few more thousand generated by people who can't write but think they can (and some of those the kind of people who don't understand what periods or paragraph breaks are for), a thousand on topics that are not of interest to you (for me, it's the kind of people who write a story with magic but think that because there's a spaceship, it must be sci-fi), all in the hope that one or two of them are actually good? You'd spend so much time reading bad stories that you might not make it to the first good one. Even editors start with a summary or an excerpt and end up skimming some things before they decide whether to take on a novel, and they're expecting to make a profit from doing it. I don't have the time to be an unpaid filtering editor when I can go buy the result of someone else's filter.
Hey, that's a popular comment. You should develop it into a story and publish it.
Seriously: I don't think a crowdsourced-edited literary journal would be much fun to read, personally. And we already have a whole bunch of them, for all sorts of genres, available online. The whole point of print journals is the triumvirate of editing: acquisitions, development, and copy. That's where the value is. Oh, some people like having it in physical form – I do myself – but the editing is mostly what people pay for.
In the '70s, for example, Analog was what it was because of Ben Bova's editorial hand and the rest of the board and team. With IA'sSFM you can see significant differences in the Scithers, Moloney, and Dozois eras. Readers come for the stories, but the stories they come for are as much the result of the editors' labor as the authors'.
Then you get into the whole business of AI-reviewerbots:
I {{adverb}{sincerity=0.9}} {{verb}{tense=past_participle}{favorable=true}{emotion=0.8}} this {{medium}{refer_subject=True}}. The author's {{adjective}{favorable=true}{sycophancy=0.75}{genre=creative}} use of {{abstract_noun}{medium=words}{floridity=0.9}} and {{adjective}{obscurity=0.833}} {{noun}{pomposity=1}} makes {{pronoun}{traditional=Null}{personal=True}{case=accusative}} {{random}{a shoo-in|short-list material|sure to be dismissed outright by the blinkered judges}} for {{noun}{category=award}{medium=literary}{susceptibility_to_bribery=1}}.
Starting a year or two ago, I noticed the emergence of a style of IT troubleshooting "how-to" web pages that had clearly scraped the technical steps from another source, and then padded it with a lot of infuriating waffle, with a real "English is my second language" vibe.
I'd originally assumed that it was just lazy wannabe tech-gurus, with nothing new to contribute, but it seems likely now that they are completely auto-generated.
They can ruin your search results with identical copies of the same obvious solution to a problem, making the truly useful and obscure nuggets of info so much harder to find.
The AI training data, if it continues to derive from the internet, will presumably increasingly contain more AI-generated text, meaning that future generations will be averages of averages until the entire internet is just a kind of linguistic grey goo scattered with SEO terms. Thus ends the first age of the internet.
> There needs to be transparency from the authors and the platforms about how these books are created or you're going to end up with a lot of low-quality books
Proof that executives of all stripes are utterly divorced from the organisation they execute. Low quality? Has this person never stepped into a bookshop?
And yet I am entirely confident that the use of souped-up predictive text to write books will manage to make the average standard worse. These toys are fun to play with but they aren't going to create anything of merit, just more averages of their inputs. If they do come out with something interesting there's a good chance it will be simple plagiarism and sooner or later people are going to get sued when GPT (or whatever) regurgitates a bunch of someone else's text into "their" book.
The low-quality stuff you can find now is at least mostly the kind of low-quality junk that can be produced by someone actually trying to write and who probably thinks they created something at least mediocre. Imagine what heights of junk can be created by someone turning on an automatic random word generator and not bothering to check the results. Imagine how much harder it will be to find something you want to read when there's ten times as much junk you have to get past in order to find it.
It could be argued that eBooks, particularly Kindles, have already done that. Certainly, I've found a lot of free/cheap works on the kindle store that would never have seen the light of day if quality was a serious factor.
But even so, this lament seems old - wasn't 'pulp fiction' the previous threat to serious writing? What we do need is for publishers to clearly and prominently state where books have been auto-generated. Then we can vote with our wallets.
Which doesn't really help publications like Clarkesworld, though. The danger is that this sort of spam might make it very difficult for smaller publishers to survive, and that would be a loss to us all.
Analog was one of the original pulp magazines (though with several titles: Astounding Stories of Super-Science to begin with) and it's still going reasonably strong. It also accepts unsolicited submissions and to be honest I'd expect them to be an obvious target for this statistical vomit. Though I wonder about the chat-gpt training: if it is informed by the quality of freely available material, it won't be too hard for the editors to tell the difference. And noting that many of the early SF stories _are_ available, it's likely that the output will show the same cultural mores as the authors of the time presented. That should work well!
I think it was Analog who published a story in which a computer took over writing for an author, getting better and better until every book published was by this 'author'... book event horizon?
Its very similar to with the advent of smartphones everyone is now a photographer. It will be the same with computer generated books just as their is an amazing photo in the torrent of dross there will be an occasional gem of a book. The major problem will be finding it.
Maybe we can all just generate our own books - no need to buy them anymore but scream loudly if you create one worth reading.
Yes and no. Serious writing is always valuable; if you read through centuries old books then you'll find that as the cost of printing was high, the quality of writing was a lot higher because editors (and the reading public) had higher expectations.
As the cost of printing has diminished, so has the level of editing. This culminates in effectively free printing and distribution (kindle store) and zero editing which results in submission quality reaching the same level. That is not per se actually a problem.
There has been, still is and almost certainly always will be a market for publishers that filter out dross and only print (or provide) books with a value in excess of their cost; the problem is simply that the cost of zero effort submissions to a publisher is non zero. This will (have to) result in increased barriers for entry; which will have the pretty inevitable effect of excluding poorer entrants who might otherwise have received some tips for improvement.
Publishers could solve the problem and continue as they are by ye olde fashioned method of asking prospective authors to drop into the office for a cup of tea and a chat when submitting their manuscripts; therefore imposing a time and travel cost on the person submitting something, or only accepting things a submission via email on a personal recommendation from another author. Or simply by charging money for reading and reviewing; which again increases the cost to everybody else other than the prats spamming people.
@Peter2
Unfortunately in paid for books editors do not always do the job they should do.
Years ago when daughter was young, partner & I read her the Harry Potter books (in smallish chunks each night) as a pre sleep activity.
You could see that in the later books the editing was minimal and they were really bloated compared to the early books (which were reasonably fast paced), a good editing session would have at least halved their size.
We assumed that, by the time of the later books, that HP was such a big phenomenon (& making so much cash) that Rowling was able to ride roughshod over the editors demands in a way that was not possible before she hit the "big time".
We did end up on the later books doing a bit of a speed read - speaking a concise version in own words to her, another pause & speed read ahead, then speak the concise version etc. If we had not daughter would have lost all interest as ratio of words to useful content was too high for a young child (probably fine for an older Potter fan reading it themselves, but no use for "bedtime story" reading)
"if you read through centuries old books then you'll find that as the cost of printing was high, the quality of writing was a lot higher because editors (and the reading public) had higher expectations."
I haven't always found that. I think at least some of that may be due to the fact that we tend to know a relatively small set of classics which were better than a lot of other books written at the same time but which didn't endure as well. Comparing the best thirty books from the 1840s, or at least a selection of thirty from the good pile, with the sum of what's available in the 2010s, it wouldn't be surprising that the 1840s look better. What I tend to find in less popular older books (and even some of the popular ones) is a lot of extra loquaciousness which certainly sounds fancy, and probably it did at the time as well, but doesn't necessarily add to the quality of the book.
It is within the capabilities of any writer, if he be endowed with a thesaurus and a dictionary, and the will to act thus to make his speech in comportment with the styles of yesteryear, or perhaps a more meritorious attempt would be to render such speech in a manner that resembles his contemporaries' imaginings of such a style without a mimesis most ardent in its honesty, to adopt as his own any oddities of vocabulary or syntax that he believes will engender the accolades of which he is desirous. It doesn't make what they have to say any better just because you have to disentangle the run-ons, or worse unnecessary references to something else, to get there. I've seen people try to write that way today. They produce terrible stories and think that sounding formal will help, but they are wrong.
Publishers could solve the problem and continue as they are by ye olde fashioned method of asking prospective authors to drop into the office for a cup of tea and a chat when submitting their manuscripts; therefore imposing a time and travel cost on the person submitting something, or only accepting things a submission via email on a personal recommendation from another author. Or simply by charging money for reading and reviewing; which again increases the cost to everybody else other than the prats spamming people.
Travel is reasonable if you're trying to get your novel published. If you're trying to get a 400-page book to market, then train tickets and a few days doing the rounds, pitching the book over a cup of tea to publishers (or indeed, soliciting agents to go and do that for you!) is quite reasonable and a small cost compared to the time you've poured into the work itself.
But this doesn't work for short stories, poetry, articles or other short-form work for monthly anthologies or magazines. Even if we accept that's just for your first-time submission, and they'll accept emailed submissions going forward, it's not practicable for most people to jump on a plane to New Jersey to meet Neil Clarke and punt a 500word short story that you might get paid £300 for (I have no idea what their going rate is, but it probably won't cover transatlantic flights and a hotel).
Also, I believe Clarkesworld haven't even provided review feedback on submissions for a while because of how many they get. Not because stuff is plagiarised, just because a portion of it doesn't make the cut - so they're definitely not going to have time to spend chatting with you over tea and biscuits. And obviously they don't want to restrict themselves to contributors in the North-East USA.
Now, in practice, it would be nice to think that you could get published in a UK magazine and then gain a reputation which would get you on the "allow-list" for other outlets. A referral from another author is a nice idea. The difficulty with that is whether - in a genre or country - how many such outlets exist. There isn't (AFAIK) really an equivalent to ClarkesWorld in the UK - we used to have New Worlds, but short-story sci-fi is a bit niche (most mainstream sci-fi magazines are focussed on news from big franchises and established authors - Dr Who or James Corey). Other comparable publications may have the same difficulty in that it's difficult/impracticable for someone in Australia or Mauritius to get on the "allow list".
Which leaves paid-for submissions & reviews. This harks uncomfortably towards the problems with academic/journal publishing (although the likes of Clarkesworld would actually be paying their reviewers!), and there's the likelihood of a rash of predatory publications popping up (as per academic publishing) where you effectively pay to get printed, or for a (possibly not-very-helpful) review. Or the "publisher" just takes your money and runs...
Interzone hasn't gone yet - it now has another owner/publisher/editor. I'm not sure if my lifetime subscription got transferred to the new owners (hmm, https://news.ansible.uk/a427.html says they emailed subscribers last month - I didn't get an email). I got bored of Interzone years ago, hopefully the new owners will make it interesting again.
if you read through centuries old books then you'll find that as the cost of printing was high, the quality of writing was a lot higher because editors (and the reading public) had higher expectations
Counterpoint: penny dreadfuls. Or children's literature of the 19th century: Geoffrey Trease, in The Thorny Paradise, talks about the fairly abysmal mass-produced work-for-hire children's novels from around the turn of the 20th century, and it was an established practice for some time then.
Maybe that's not far enough back; but the fact is that much of what survives from earlier periods is the good stuff, because that's what people preserved. So there's quite a bit of selection bias here.
Obviously in Europe the Bible was widely printed, and the writing in that is a mixed bag, at best. Malleus Maleficarum had 28 editions in the 15th and 16th centuries, and it's not what I'd call deathless prose. Certainly there was great writing that became very popular: Don Quixote and Tristram Shandy are good European examples. But printing-press economics flattened out quickly: it's estimated there were 150 million books in print in Europe by 1600. That's a lot of books. They probably weren't all great.
> ... always will be a market for publishers that filter out dross ...
Sadly, as has been reported in this very organ, there are "publishers" who effectively filter out quality. Fake academic journals and the like...
the lament is, indeed, old, and looks very much like the moment print press entered the business realm and made all those monks jobless in no time (I dread to think what trade they turned to at that point). But somehow, I feel that the analogy is superfluous. Yes, AI, or whatever it really is, is going to generate million times as much text (and soon movies, and music, etc.) as the wetware-based humans, and sooner rather than later, AI output will match and then outdo humans'. But on top of that no-cost sellable crap is unemployed and unemployable milions in all worlds, from 1st to 3rd (and beyond). And those unemployed, before they die of hunger, might want to riot and hang or string a few of those they deem responsible for their fate. So, to end my long and pointless comment, perhaps it's time to invest in some wheeled AI-robo tanks to keep the world order and the elites safe?
"But even so, this lament seems old - wasn't 'pulp fiction' the previous threat to serious writing?"
The printing press? That will ruin the value of REAL books, wrought by the hand of a scribe!
Yeah, tech causes ructions when it first appears, then everything settles down again. Artists and writers are the current "target" for automation and many will probably drop out or stop being profitable and therefore less productive. It'll be harder, or at least different, for up-and-comg-coming artists and writers to break through and be noticed. But not impossible. We just have to look at every "craft" ever that has been partially ousted by automation or mass manufacture. Baking bread, making pottery, cabinet making, and pretty much every other product that most people buy today was an "artisan" product in the past that was hard to get a hold of and very expensive. But most of not all of those artisan crafts are still around, still expensive and still hard to get. But there's cheap mass produced stuff always available cheaply for the..err..masses. Just like "pulp fiction" and the "penny dreadfuls" of yesteryore were often very poor but sold in their droves.
That Wikipedia link holds true terror: the page count went up from the original 23 'zine pages to 76 in book form!
Perhaps the extra pages are accounted for by large print (to read through the tears) and blank pages (to allow a few gasping, untortured, breaths).
The whole of society has simply become a consumer based side of the equation, no one wants to learn and practice how to be creative. Photography, like many creative artforms, is not just an action ( clicking a button ) it takes hundreds of hours of practice to be able to spot good opportunities that will make intriguing images, this is the joy. No one sees this, they simply think the camera does all the work!
All we will get from AI art, AI photos and AI writing is just a glut of cookie-cutter, second rate garbage being churned out. This is what will happen with this lazy, "get a machine to do it" attitude.
We will simply be a bunch of lazy, stupid, feckless morons who will not understand or appreciate the joy and craft of creativity. What am I say? We already are!
I can remember a craft fair I went to. There was a guy selling stunning photos printed onto local Caithness slate. One twonky asked about the pixel count of the photographers camera and when told mocked it because his smartphone had a higher pixel count. Smartphone ... Stupiduser
I don't know your environment, but around here nearly everyone is creative. If they're not artists, they're artisans; if they're not either of those, they're spinning stories to friends and family. There are plenty of people who devote much of their energies to being creative in one form or another.
That definitely doesn't square with the creative people I know, many of whom have years of experience. But from the perspective of someone trying to earn a living, it does get increasingly hard to justify putting a lot of blood, sweat, and tears into your creative output when someone can do it faster and "good enough" with automated tools. Cheap, mediocre goods will always push quality goods out of the market; that's basic capitalism.
Make people sign a contract when submitting a story, guaranteeing that it is all their own work and contains no computer generated text. The contract should stipulate that if a text is found to be generated by an AI, the submitter is liable to pay $50,000 in compensation for wasting the editor's time.
That should vastly reduce the botrot and encourage more original and interesting stories.
"and contains no computer generated text"
Better not to leave that quite so vague, or anybody's spell check could fail, depending upon how literally you interpret "computer generated".
Otherwise, have an upvote. I'd be quite happy to sign such a contract. My various stories might be crap, but they are one hundred percent human generated crap.
At least get the quote right its "I'm sorry Dave, I'm afraid i can t do that"
When developing an audio editing system in the 90's the head of the team was called Dave. For a bit of a laugh one of the software engineers programmed every key on a development system to play the clip.
A comment above by Little Mouse provides a possible answer or at least explanation to a situation I am suffering from.
For my sins I have the honour or misfortune of being a senior moderator on bleepingcomputer.com and for a little while now I have been bothered by a not large but not insignificant number of posts which are suspected of being straight plagiarism but it is difficult finding an original source for many of them. His description -" "how-to" web pages that had clearly scraped the technical steps from another source, and then padded it with a lot of infuriating waffle, with a real "English is my second language" vibe." fits what is happening to a T. This is something I think that I and my colleagues are going to have to discuss.
Chris Cosgrove
I hate to think of the hours I've spent looking for an actual answer to a Windows problem wading through webpages all giving the same answer that just doen't work, or is something any fule knows. Not quite as basic as turn off, wait, turn on - woops having written that some of them ARE that basic, in fact that is the recommendation.
The fight against computer generated images and writing my just be beginning, but it's been prevalent in music for a while now. A majority of music you hear is arranged, produced, and played by computers from a prefabbed batch of samples. Especially, in the film and video game world. Nobody is playing anything anymore, It's all pressed out in mass by computers. That's why nobody is listening anymore, and it all sounds the same. We are being educated and entertained by computers at this point. There is a saying in electronic music that has been around for the last 25 years.
"We used to play machines, now they play us."
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
