• Hi Guest!

    The costs of running this forum are covered by Sea Lion Press. If you'd like to help support the company and the forum, visit patreon.com/sealionpress

ChatGPT and creating alternate history

I'll say (though biased as its my current obsession) that AI art has more value in online AH storytelling. The use of it to replace stock images can carry over into a subgenre filled with them. It can also make portraits of distinct fictional people, which will help....

...if you're willing to put in the energy and effort. To get a really good AI picture means a not insignificant amount of effort and manual work. And I don't see most people who just plop down vaguely related stock photos as wanting to do that.

Which is why I feel it'd mostly be a small step up, from "stock photos" to "obvious low effort AI pictures".
 
I haven't looked at a Corbis TL in like a decade so it makes no difference to me if the people writing those replace pictures of real no-name American pols with AI pictures of fictional American pols. The thing about AI art is that it's gotten to the point where it can pass convincingly as something a human's done and when it screws up that still produces something visually interesting and that's pretty cool. ChatGPT can do some convincing generic thinkpieces and wiki entries but not fiction yet. Might have an opinion when it can.
 
The thing which interests me most about the current state of AI is that there are also ongoing efforts to write software to detect what's written by AI. So in other words, there are AIs learning to write like a human (for a given value of like), and there are other AIs trying to learn how to tell an AI from a human. An AI arms race, as it were...

... and according to the current model of AI detection software, my writing sample means I am "probably human." I'm not sure whether I should believe them.
 
Coiler said:
Fanfic writers are already using chatbots to aid them in writing, and other fanfic writers are screaming bloody murder at them about how they are insulting art with their soulless heretical abominable intelligence machines.
Click to expand...
That makes sense to me. I must admit I find the way the issue has become emotionally charged baffling. Which is weird because I'm usually pretty anti-tech. Like, I'm anti-nuclear power, HS2, etc but the harsh anti-AI response seems odd to me.

I think part of it is that right now, AI is strongly associated with the hard right and also it's another tech bro concept - and tech bros have historically low levels of trust after spending a decade pretending that crypto and the metaverse will be the next big thing. This means we're primed to see AI as bad, and also liable to be ineffective and overhyped.

But the thing is, AI tools are already more widely adopted than any of the shit boring men have been shilling for a decade and it seems pretty evident we've not seen full adoption yet. We're in a weird stage where people online will tell you that AI will never be able to draw a picture of a handshake while doctors are already using it to diagnose lung cancer. That said, I don't think AI will be able to write AH well any time soon because there's such a small data set for it to work with and it's a genre with very strict formal rules around causation and genre conventions. AI is useful for art and organising data, but I don't think it will write an interesting AH scenario any time soon.

I don't think technology implementation is inevitable. I was at a talk about AI yesterday and one of the speakers made the point that its technology that's being developed by exploitative land and labour practices and resistance is possible. I think it is viable and desirable to resist big tech in online spaces. I know some smaller story writing collectives where AI is very much not done and doing it feels against the whole ethos of the thing. With careful moderation they can keep it at bay. However, SLP and even AH.com are just a tiny bit of the AH community and I don't think we've really properly faced up to how uninfluential the online AH community is to the broader creation of AH. Even within online AH communities there isn't the ideological uniformity needed to really resist creators implementing AI. I don't think it will be possible to lockdown the genre so AI doesn't happen in alternate history circles.

I think AI represents a very different model of online content creation and consumption. It's a move away from craft and I do think there'll be a resistance to that. I feel like the art of resistance to AI and big tech could open some interesting new options for AH creators. There are a lot of very political people in this genre so its a rich area for exploration.

So basically, we'll be less impacted by AI that the big online fandoms. But less able to resist it than others. But the nature of our community creates the possibility of AH playing a key role in pro-human craft movements that develop in response and resistance to AI.
 
Sideways said:
That makes sense to me. I must admit I find the way the issue has become emotionally charged baffling. Which is weird because I'm usually pretty anti-tech. Like, I'm anti-nuclear power, HS2, etc but the harsh anti-AI response seems odd to me.

I think part of it is that right now, AI is strongly associated with the hard right and also it's another tech bro concept - and tech bros have historically low levels of trust after spending a decade pretending that crypto and the metaverse will be the next big thing. This means we're primed to see AI as bad, and also liable to be ineffective and overhyped.
Click to expand...
I think you’ve answered your own question here. The reason why ChatBots and the like create so emotion from writers and creatives is that it is for the most part they are being sold by the people creating them as a means of eliminating writers and creatives and their subjectivity and demands for better pay and working conditions. When someone shows you who they are, believe them.

Beyond obvious self-interest the idea that an AI’s work can just replace an actual writer’s betrays a pretty miserable view of creativity (and, as I keep saying, ChatGPT can’t actually create and explicitly doesn’t claim to). That it’s just words in order to amuse us and nothing more. Dismissing the human factor when what makes art art is the craft, the fact that an actual person did this with their mind.

That last point might seem over the top English literature graduate pretension but it’s pretty self-evident in my opinion. Why else would soulless corporations spend so much money on behind the scenes documentaries and press tours where writers, directors and actors talk about how much of their own heart and soul they put into Superhero Schlock Seven? Not sure we’d get as much press coverage of the computer programmers talking about what coding they used to better dredge through pre-existing scripts.

My view on ChatGPT is that while a lot of the technology behind it might prove to be transformative in the long term, at this stage chatbots are a proof-of-concept and exactly nothing more. And that goes for ChatBot-produced AH as well.
 
A question I was drilled into a lot doing my MA and which is one that I ask myself whenever I do sit down to write is "so what?" and variations thereof.

A robot writes a TL. So what? Beyond the novelty of it, what else is there? Why should I care? Is there something going on compositionally, structurally, emotionally, anything beyond the simple fact that it's a robot vomiting its soulless imitation of cut-up exercise, that justifies reading? No? So what?
 
I can say that the erratic Google Bard can write AH: I asked it to list people successful in both boxing and MMA, and one of the responses didn't involve Holly Holm (far and away the most successful and prominent of that nature) once.
 
I've found ChatGPT and Bard both pretty useless at writing the history itself and I haven't used it for that at all, it does to me seem to take out the joy of the thing. Art without an artist isn't really art at all.

BUT I have found it helpful for my TL/Game on the other place, which revolves around a Scandinavian influenced England, for sorting out the conlang. Scouring Old English Dictionaries was fine but time consuming and for rarer words the dictionary obviously can't say "There's no word for that but this is close enough" Just being able to say "What is [WORD] in Old English" and get a coherent answer has been very useful. It does sometimes lie/make stuff up when it doesn't know and can be inconsistent but I always verify with Wiktionary as well as other sources and this has made this process pretty painless and one of my favourite parts of writing the TL. It can also sometimes be useful in understanding certain concepts, particularly the development of certain technologies (what the prerequisites can be) or generating names, places and the like.

Some of my players have also used AI to generate images for ITTL characters, battles, buildings etc which is useful. Ethics here are complex but if its pulling from national portrait galleries and the like I think combining various faces into one is no less immoral than using a historical portrait or painting in an AH context as a fictional character. As it gets better I think this is probably one of the better uses of AI for Alt History - if the ethical issues can be resolved/settled.

I'm very AI sceptic and absolutely hate most full AI media I encounter however I do think it can have a role to play as a sort of beefed up search engine.
 
Well, that's it, guess I'll just pack it up
ChatGPT said:
It's difficult to predict what America's top food would be in a hypothetical socialist world, as there are many factors that could influence food culture in such a society. However, it is possible to speculate based on some of the principles of socialist ideology.

One potential outcome is that the focus on communal ownership and equal distribution of resources might lead to a shift towards plant-based diets, as animal agriculture is often criticized for its negative environmental impact and resource-intensive nature. Therefore, plant-based dishes such as vegan burgers, tofu stir-fry, and lentil soup might become more popular.

On the other hand, socialist ideology also values diversity and cultural expression, so it's possible that traditional American foods could still be celebrated and enjoyed. However, in a more equitable society, access to quality ingredients and cooking equipment would likely be more widespread, so even simple dishes like fried chicken or mac and cheese could be elevated to new heights.

Ultimately, it's impossible to predict exactly what America's top food would be in a socialist world, as food culture is influenced by a wide range of social, economic, and political factors.
Click to expand...
 
Last edited:
JesterBL said:
A search engine that you have to fact check with a different engine because it will confidently state stuff it just kinda made up is not exactly a strong use case for something.
Click to expand...
Ah but the thing is if I search it on Bard or ChatGPT it always gives me something and 90% of the time its accurate. A lot of OE (and other rare language) dictionaries online are really really lacking and many that don't include a dedicated translation service can be achingly difficult to trawl through if you don't know where you're starting from for a word's origins. Searching manually can take forever and checking is literally as easy as putting the suggested word into say wiktionary. Trust me I have spent probably tens of hours doing this, it's made life easier. I'm not here to preach for AI as some wonderous new world, I think its at best to be used sparingly and it'll probably put me out of a job in the next few years (researcher for a newspaper) but there are clearly some cases where its useful.
 
Major Crimson said:
Ah but the thing is if I search it on Bard or ChatGPT it always gives me something and 90% of the time its accurate.
Click to expand...

Until it gets trained on more garbage data produced by AI then that 90% drops to 81% to 65% to 43% until we're eventually just in a dead internet scenario of chat bots making up nonsense for each other.

And no, a tool (for a very simple query of x(english) = x(old-english) being right 90% of the time and confident 100% of the time and impossible to discern which one it is without conferring with other resources is not useful. It's gonna produce garbage.

AI is at best a joke and at worse a threat. It's a solution in search of a problem.
 
The issue is that you're not actually describing a real use case- the value proposition you've put up is that AI can be a useful replacement for BehindTheName.com, which we already have and is basically as effective as what you just described.

I'm not really emotional on this topic (my livelihood is not threatened by AI, I will admit freely), I just really don't think that AI like Bard or ChatGPT has shown any utility, especially not for research. At this point it makes memes and crypto-influencer posts and sure, it can probably spin up lists of fantasy names with the best of them.
 
I don't believe ChatGPT or any of the other current offerings are anywhere near real AI. On the other hand, I have no doubt that it is coming. Will we see anything like the Minds of Ian M Banks's Culture? I don't know, but I think it's probably a lot closer than Cold Fusion.

So why all the current hype? Charles Stross has one suggestion
 
JesterBL said:
Well, that's it, guess I'll just pack it up
Click to expand...
Chat GPT said made up:
On the other hand, socialist ideology also values diversity and cultural expression
Click to expand...

Morales and Lula kind of like this, but Stalin is astonished, Sukarno is angry, Mengistu is apoplectic.


I can only assume that this is picking up on socialism = all things (some of the) US right hates?
 
You must log in or register to reply here.
Back
Top