Artificial Intelligence and Creative Works?

General Off-Topic Discussion
1

I saw this post from Clarkesworld today:

https://twitter.com/clarkesworld/status/1627711728245960704

Apparently they’ve been inundated with artificial intelligence (presumably ChatGPT-based) submissions that they’re having trouble filtering out quickly, and can’t get through their submissions any longer. This is a real shame, given CW’s stance as a publication that pays, responds quickly, and has generally been a haven for new talent.

I can’t help but to keep thinking about how things are changing, and what my fears are.

Partially that’s because I’m hoping to query a novel to agents later this year, and fear I’ll be crowded out. Partially it’s a broader fear of what systems like this will mean for creative works and fields more generally.

It also make me wonder if we’ll need to start implementing anti-AI policies in comps, or if there will be some comps focused solely on AI creations.

I’m curious to hear where people sit with this. I know in at least one or two other threads, people have started to grapple with it a bit. But it seems mostly superficial discussion so far.

Not sure where else to leave it off, but I know there are a lot of smart people here with more historical context and experience than me!

12 Likes
2

I’m sure this forum doesn’t need to see go on another unhinged-sounding rant, so I won’t go into too much detail.

On one hand, it’s extremely fascinating, and I’m super excited to see all the crazy things this technology will do in the future.

On the other hand, it literally represents the end of an era, and I don’t know what the next era will look like, and I also can’t help but feel like this will make my years-long and desperate fight to start a software engineering career even harder.

Mostly, it will probably remove opportunities for new engineers to flex what they can do, because once these tools start really digging into the software fields, a lot of engineering is going to be for systems instead of actual code. I suspect most hiring teams aren’t going to have a way to effectively evaluate new engineers, because—traditionally—that had to do with the applicant’s coding portfolio and background.

I have the skillset and knowledge necessary to engineer systems in a post-AI software field. However, a lot of the usual interview questions will be obsolete, once this technology takes off, and the industry will need to discover a completely new hiring process, and it’s gonna make the cyclic problem of need-experience-to-get-experience even worse.

I suspect this is going to make my immediate survival chances a lot slimmer in the next 5 years, as I am someone who is literally only good for software engineering, because of disability.

So…y’know…existential dread aside, and specifically avoiding wild tangents from last time, I can’t wait to see what’s over the horizon! :grin:

12 Likes
3

I can understand someone seeing the submitting of an AI generated story as an easier (less effort on the “writers” part) means to get published and make money, assuming that it gets accepted.

But I think the acceptance of such works, assuming that it’s possible to tell AI from non-AI (a), is potentially the start of not needing Authors or at least not as many. As the “publisher” is just as capable of entering a series of words into an AI’s prompt as an Author is, so what exactly do they need the Author for?

(a) The measure of “success” for an AI generated work would be how well written it is. So if it is as “successful” as a well written work by an experienced Author then how do you tell it was generated by an AI. Short from getting the Author to some how “prove” they wrote it unassisted.

5 Likes
4

I saw that article a couple of days ago.

In this case, it’s clearly an attempt to scam some money out of a magazine without doing any work.

It’s a problem, but it also doesn’t speak to how the creative arts will deal with actual attempts to do creative work with AI assistance.

(For our IF world, I am actually more worried about AI-generated spam appearing on the forum, in IFDB, on IFWiki, and so on. That will be polluting history, not creative work.)

11 Likes
5

What an absolute nightmare to hear about the troubles Clarkesworld has been facing. Super sad about it! Fingers crossed there’s some sort of resolution soon- I’ve heard wonderful things about them in the more writerly circles I dabble in, and it would be such a loss if they shuttered their doors permanently.

7 Likes
6

Software Engineering encompasses such a diverse field, that you have great opportunities. Can you do YouTube? Can you showcase programs of your designs? If yes, then create your own luck and be a consultant. Or maybe a teacher? On-line classes and all that.

Certainly no one knows everything, and opportunities for collaboration exist. If you can design AI, even rudimentary ones, you can built some pattern recognition programs for Face Detection, Sorting Color/Shapes, OCR, Stock Market.

Or maybe design your own Twitter software. I hear Elon Musk is looking for a new CEO. :grin:

4 Likes
7

This is the standout part of your post to me. CONGRATULATIONS! Writing a novel is a huge accomplishment. Can I tempt you into saying a little bit about it?

I think we’ll weather the invasion of the AIs as well as we’ve weathered all tech up til now: which is to say, we’ll flub it and the world will be a worse place. I suspect that it will go much as social media and internet have gone: capable of immense good, but we’ll use for cheating in school, making elaborate fart jokes, and screwing with people we disagree with. Same old.

14 Likes
8

I imagine it will make it easier to automate disinformation campaigns on social media and other “fast” content sources.

Our own futzing around with it is all but certainly supplying test data.

It’s worth asking why these breakthrough technologies are purposed for content creation. The labor that they might one day displace is not terribly expensive–not compared to the tech, anyway. Lord knows we don’t pay writers and artists enough.

People don’t spend billions of dollars just to see what happens or spoof Inform code, or because they love the written word.

8 Likes
9

I can offer some general observations about the impact of the AI technologies upon society. People have been fearful of computers for a long, long time. Witness the several episodes of the original Star Trek series in which computers threaten to take over. This computer-phobia has shown up in TV shows, novels, and movies for decades. The current round of hysteria about computer technology is based on more concrete issues – specifically the use of computers to generate text, software, and imagery that appears to rival what humans can do.

Never forget that the current LLM technologies are nothing more than parrots. They have a huge database that they plumb for content matching whatever input they get. If you input “The Ides of March are come.” it will surely respond with “Aye, Caesar, but not gone.” That doesn’t make Shakespeare obsolete. The LLM is merely parroting Shakespeare.

These models do threaten some people: the parrots. You know the people I’m talking about; they’re the hacks who merely regurgitate stuff they’ve seen elsewhere. They slap together a plucky young hero, a beautiful damsel in distress, a few dragons, a wizard here, a troll there, a magic sword, and call it a story. These people are definitely in trouble, because any LLM can do the same thing.

Power technology, whether based on steam, coal, oil, or electricity, rendered the use of human muscles obsolete. Societies were upended because muscle power had always been of overwhelming social and economic importance. Millions and millions of people whose livelihoods came from the use of their muscles were rendered obsolete. People just had to rely more on their brains than their muscles. Isn’t this better? Would you rather spend your life handling bales of hay?

This AI technology is pushing us further up the scale. Hacks who just parrot stuff will be driven out of business, and success will go to the more creative people. Wouldn’t you rather live in a society where human creativity is economically important?

4 Likes
10

“Just become Shakespeare to remain economically relevant” sounds positively dystopian.

10 Likes
11

The issue I see is that your description of a “parrot” Author also describes in general terms:

  1. people trying to get started in the story writing industry who haven’t had enough experience yet to craft a “more original” story.
  2. many currently selling authors who base their story lines on the any of The Seven Basic Plots.

eg. How often does a truly original story line or plot get written?

7 Likes
12

Having tested out various AI writing sites, I think the current state of AI text generation is incapable of creating anything good. It can make produce short snippets of some quality but anything long becomes nonsensical, and it can’t remember basic details. A good short story is beyond its capabilities. The major use of AI in writing, from what I can tell, is stuff like spam, editing for basic grammar, and throwing together bland articles on stock market price shifts. Just generally doing the boring stuff nobody is interested in.

But as the technology improves, I can see things changing in the writing sphere the way AI-generated art has thrown internet art communities into disarray. AI-generated art has gotten good enough that genuinely talented artists are being thrown aside in favor of algorithms trained on their work and the work of others like them. Similarly, more and more genuinely talented writers could lose the opportunity to get recognized as people turn towards algorithmically generated content. The technology is promising, and there are plenty of exciting things that can be done with it, but the consequences as it improves will be severe. And I don’t see it not improving. People are throwing billions of dollars into this. Most of that AI money goes to already-rich technocrats and companies as well, instead of actual artists, who aren’t exactly known for being well-off.

I’m not a professional writer in any way, so I could be wrong about the effect it will have. But there’s my two cents.

2 Likes
13

I don’t think I have anything new to say on the negative possibilities front.

On the unanticipated positives front, of a sort, I saw a post on facebook by Robert Henke, creator of Ableton Live (digital audio software, one of the two main ones I use, the other being Apple’s Logic) and whose artist guise is Monolake, which speculated the following, and did get me thinking:

"January 29 at 10:42 PM - Sunday morning thoughts:

AI is perfect in re-creating existing things in a highly believable fashion, be it images, text or sound. Even the most outworld-ish visual creations and mashups are based on known visual tropes, this is what makes it so powerful. We see new things that are not really new. It is all HR Giger or Rembrandt or Studio Ghibli or Wim Wenders, etc…

( No, i am not dismissing AI as tool here, there are fantastic artists out there using AI, but their process is typically highly iterative and driven by a vision that goes beyond of 'how to make an image look like XYZ (which now ironically also can mean, like the stuff AI artist XYZ is doing 😉 ) )

However, if AI for audio would have been ready before the invention of UK dubstep, I doubt it would have been able to create a ‘Burial’ or ‘Code 9’ sounding result. And even if one of the random seeded creations would have sounded a bit like it, it would not have been recognized as important, since the social interaction and resonance that is required for invention would not have been there.

AI in sound forces creative people to rethink their practice: If you cannot contribute anything that is highly personal and unique, the AI will do it better.

There is no point anymore in ‘how to produce a professional dance track’ tutorials.

The more i think of it, the more i find this comforting. Don’t compare, do what you really really want to do. Leave the rest to the AI.

for reference, look e.g. at this: https://google-research.github.io/seanet/musiclm/examples/"

-Wade

4 Likes
14

I’m of the opinion AI is going to present a problem primarily in the way it relates to labor and the way people value it. It’ll likely get better and eventually stop having most of the goofy flaws it has now, but it’ll never be able to truly replace human creativity, because creative endeavors are fundamentally communicative and it’s impossible to completely remove living, breathing people from that context. That said, I think it’ll be a cash-cow for a creatively bereft management class that traditionally doesn’t value soft skills.

It’s kind of an “old man yells at cloud” thing for me, but I think it’s the inevitable endpoint of the current framing of most creative endeavors as “content.” The internet is organized around various platforms that need to be filled with stuff, and we’ve collectively accepted this idea that we’re all supposed to make stuff to fill them, and if they can use technology to make the stuff quicker and cheaper, that’s what they’ll do. It’d be a good thing if it was just going to free people up from doing menial tasks, but historically it just means new menial tasks get made up for people to do.

So, in short: I think AI is going to make a lot of things aesthetically and creatively worse, produce mountains of digital garbage, and put a lot of people out of work, but the kinds of practical/technical problems Clarkesworld is facing will likely get resolved and there will always be a distinction for creative work that actually represents communication between people.

5 Likes
15

That’s very kind! I really do appreciate reading everyone’s thoughts in this thread.

Here’s the pitch: it’s an adult fantasy novel where magic is a soft skill, theoretically learnable by anyone, but which also requires a great deal of effort for very small returns. Few people are moved by the thought.

Our protagonist is on the verge of failing out of a renowned institution for conjury when two physics-defying natural disasters strike - first at the capital, and then over a small port town not two weeks later. With most of the able conjurers off at the capital, the protagonist is sent with an unwilling mentor to investigate the latter, a mile-high thunderstorm which refuses to move or relent. As the pair dig deeper they chase mysterious figures of ambiguous loyalty and uncover truths about the nature of magic itself.

I finished the first draft a couple of weeks ago, and will begin editing once I’m a bit more squared away with Spring Thing. I’d never written a Man v. Nature story so it’s been quite a fun journey, and even if it goes nowhere I feel I’ve learned a lot for the next story!

9 Likes
16

You don’t have to be as good as Shakespeare, nor must you create grandly creative stories. You need only be better than a parrot.

Might I suggest that we underestimate the intellectual effort that goes into writing a story? Let’s think about the matter by speculating about the algorithms used by the human and the AI:

The human establishes a theme, how that theme might be expressed through a conflict, creates a protagonist and an antagonist to carry out that conflict, then figures out an introduction to the characters, how the conflict slowly develops, reaches a crisis, a resolution, and a denouement. Think about that process; consider the intellectual energy that must be brought to bear on these issues.

The computer scans a vast array of text looking for tags matching the components of the prompt and assembles its findings in no particular order. It could easily have Darth Vader slice Luke Skywalker in half at the outset, then have Luke at the farm pining for adventure while Leia runs through the Death Star, before concluding with Han Solo shooting C3PO at the bar.

Do not underestimate the complexity of our stories. Here’s a standard routine I often use in my lectures:

Here are two stories:

[in falsetto]
Itsy bitsy spider crawled up the gutter spout.
Down came the rain, and washed the spider out.
Out came the sun, and dried up all the rain.
And the itsy bitsy spider crawled up the spout again.

This is a complete story in just four sentences. It has a protagonist, an antagonist, a crisis, a satisfying ending, and even an edifying moral lesson. Any five year old child can understand and appreciate this story.

[aside: one of my greatest regrets is that I did not use this routine in my invited lecture at the Sydney Opera House, thereby depriving myself of the opportunity to rightly boast that I had sung by invitation at the Sydney Opera House.]

Story #2: Once upon a time there was a handsome young prince who lived in a shining castle atop a hill. One day he leapt atop his snow-white charger, galloped across the cobblestones in the courtyard, out the drawbridge, and down into the valley, where they fell into a hole and they both died.

Any five year old child can recognize that this is NOT a valid story.

Challenge: design an algorithm capable of determining which of these two stories is valid and which is not.

Not possible.

2 Likes
17

You people better shut up and get back to work. and work harder. and forget about that pay-rise, because AI’s going to take your job!

How many times more am i going to hear this BS. Before it was 2008 layoffs, before that it was globalisation, the dot-com boom etc. etc.

I agree with @TheBadChrisCrawford, AI isn’t going to create anything new, discover anything or come up with new ideas.

Show me the self-driving cars! Where are they? Not so long ago, all you bus drivers, cab drivers and delivery trucks were to be out of business. Now there is silence.

Yes i know, AI is getting better all the time. It’s early days yet. It just needs to be scaled up. Computers are getting faster all the time.

I’ve also been hearing this excuse since the 1980s.

The truth is, the current concept of AI does not work like the human brain. They like to pretend that it does and they liking using the same terminology. But it isn’t the same. Instead it’s just a giant pattern-matching engine. In that capacity it works well; recognising things works.

But, AI is not capable of:

  • maintaining a context
  • memory
  • reasoning
  • learning (outside of its training set).
  • adaptation
  • original ideas.
  • thinking
  • problem solving

That’s my $0.02.

1 Like
18

That sounds suspiciously like Clip Art collections. A lot of artist lost their job once cheap arts becomes available. Then the market changes, and the artists evolve to a higher standard. Can I say Adobe benefited tremendously? This is about the time they stopped selling programs and go with subscription model.

Artists are still around. Clip arts are still around. However, looking back, it was mostly various buttons and borders. Nowadays, we have really high quality photos. AI will mean we’ll have high quality paintings, as well.

Part of the consideration is the cost. There’s so much computing resources used, that it’s cost prohibitive without time share model. Of course, that’s what internet is, so there’s that. A skillful artist can do it quickly, however, so it boils down to:

  1. Do lots of prompts and see which is best
  2. Do it once by hand, and do it perfectly the first time

My preference is number 2, but I’d be lying if I say that I never gave number 1 a try. I guess what I’m saying is: people evolve. They’ve always been, and they’ll always do. Life goes on…

2 Likes
19

The “great deal of effort” sounds a little like what’s required to do “magic” in David Brin’s The Practice Effect, which I found very enjoyable.

I also like the Man v. Nature basic premise and look forward to reading the story when you’re ready to stop the journey on this one. Hopefully I see the notice of its release.

4 Likes
20

You’re splitting the syllogism. Or at least hiding part of the argument in that “better than”.

Consider the sex life of the Australian jewel beetle. You might naively expect that all a lady jewel beetle needs to do in order to be reproductively successful is to be marginally more attractive, in jewel beetle terms, than the other nearby lady beetles. But for some time their major competition was…beer bottles. Beer bottles whose stippled, hard, shiny bottoms were far more intoxicatingly enticing to male jewel beetles than anything ordinary lady jewel beetles could offer.

Which is to say it is not just that the artist needs to be better than the putative “parrot”. We also require that the audience is sufficiently able to distinguish between “real artists” and “AI parrots”. And, implicitly, to prefer the former to the latter. Where here “prefer” is not just an aesthetic judgement—in order to remain economically viable, the “real” artist needs for his quanta of “better-than-parrot-ness” to be valued beyond the differences in speed and cost (for example) that separate their work from an AI’s.

It isn’t clear that any single element of that logical chain is necessarily true (or destined to remain true), much less that all of them are collectively true under composition. I mean what appears to be true is that actual real genius isn’t consistently any guarantee of economic relevance. Think of a famous painting, not the Mona Lisa. Some large plurality of readers just thought of Vermeer’s Girl with a Pearl Earring. Another large plurality just thought of van Gogh’s Starry Night. Neither Vermeer nor van Gogh were financially successful as artists.

Earlier you asked “Wouldn’t you rather live in a society where human creativity is economically important?” I suppose my answer is that I’d rather live in a society in which we didn’t evaluate humans in terms of economic “importance”. But leaving that aside…sure, I get what you’re trying to say. But insofar as I think I understand what you mean by the terms, that doesn’t appear to be true now, and I don’t see any way in which AI art makes it any more so, and I actively and strenuously disbelieve the proposition that the locus of decision in the matter lies in some foggy notion of an artist’s “originality” (or whatever you want to call not-parrot-ness).

11 Likes