Say, within the next 5 years?  Will it put 90% of the RPG content creators, out of work?  Will only the very best, still be able to make a living?

I think, from what I have seen, the small individual creators that shy away from AI art and writing will still be around and doing their thing. WotC will probably move to AI content full-time so that Hasbro has to pay as few humans as possible to produce D&D in the future.

We had this tread earlier this year and my opinion hasn't changed.  Here's the highlights:

1.The market is going to get flooded with low quality PWYW material that people will get sick of very quickly.  The biggest effect is that sites like DriveThruRPG and itch.io are going to get very difficult to search for good new content.

2. Established high quality content creators are actually see an up swell while everyone else is going to have a hard time getting a break-out moment.

3. Either the current retailers are going to get their act together and filter out the AI noise or they will become obsolete as someone else offers a "no AI" service for supplying RPG content.

4. Creators are going to learn how to use AI to improve work flow and it will be part of the process just like books aren't done on a type writer anymore.  AI won't replace human creators but augment them.

tl;dr Big upset for a short time while people grapple with it and then it will be a non-threat.

I mean, if we pretend that AI is frozen where it is right now- and not, say, where it was two years ago- then it will just be a task of filtering out the shitty low effort AI content and then there will be some vague semi-war between zealots and real creators who use AI sparingly and appropriately.  So I agree with BadApple if it plays out like this.

But, the question was about the next five years.  Five years ago, the ability of AI to generate anything useful was nil- people would openly speculate it would be in the realm of General AI, which we definitely still do not have anything close to yet.  Instead, yet another field- bad writing with random ideas inserted- has fallen to AI.  That seems a bit laughable and perhaps no threat to any serious writer or any professional at all, but lets be real- everyone knows that machine learning isn't done.

I have no idea what it will look like in 5 years.

Quote from: Venka on December 28, 2023, 06:02:47 PM
I mean, if we pretend that AI is frozen where it is right now- and not, say, where it was two years ago- then it will just be a task of filtering out the shitty low effort AI content and then there will be some vague semi-war between zealots and real creators who use AI sparingly and appropriately.  So I agree with BadApple if it plays out like this.

But, the question was about the next five years.  Five years ago, the ability of AI to generate anything useful was nil- people would openly speculate it would be in the realm of General AI, which we definitely still do not have anything close to yet.  Instead, yet another field- bad writing with random ideas inserted- has fallen to AI.  That seems a bit laughable and perhaps no threat to any serious writer or any professional at all, but lets be real- everyone knows that machine learning isn't done.

I have no idea what it will look like in 5 years.

I agree that just about anything can happen in 5 years.  I also agree that AI will get better.

There is a major innovation that needs to take place before AI truly can write OC.  Currently, it's just a way of rolling on really big random tables and there currently isn't any foreseeable change to that.  The developmental track it's currently on will be even bigger tables and a self checking feature to improve output.  Then again, humans were terrestrial creatures before the Wright brothers decided to upset everything.

God only knows.

something about AI, and I'm only talking in regards to AI art being used in RPGs.  Some oppose the use of AI art in any circumstance, because it takes the livelihood away from an artist.  But what about an indy designer, a one-man DIY show, who would like quality art, but never would even have the funds to hire a professional artist in the first place, because the investment-to-profit ratio would make it impractical to even have paid art?  This person now has a tool to add aesthetics to a game that never could have in the past.  It's not a threat to visual artists, and it gives the little guy a step up.

I use AI to a degree within a limited extent, AI is very good at generating a lot of ideas quickly, but they need to be filtered and refined by an actual human to be good (so far)

I'm hoping the Human Machine teaming meta is still relevant within the next decade, currently Machine + Human is the best strat, and produces the best results.

correct me if I'm wrong, but two interesting things happened in the world of AI in relation to chess, being a human managed to beat an AI in chess for the first time in awhile due to the AI being super specialized at beating other AIs which it based it's training data on, but wasn't good at predicting human reactions. in honesty this was kind of a fluke just due to the nature of the event.

the other was it's be pretty proven Human assisted AI vs pure AI tend to win at chess. I'd like to believe on some level there is something about organic life that AI simply cannot copy or emulate as a fault with it's substrate, though I admit this might be naive thinking.

Quote from: Socratic-DM on December 28, 2023, 07:38:21 PM
I use AI to a degree within a limited extent, AI is very good at generating a lot of ideas quickly, but they need to be filtered and refined by an actual human to be good (so far)

I'm hoping the Human Machine teaming meta is still relevant within the next decade, currently Machine + Human is the best strat, and produces the best results.

correct me if I'm wrong, but two interesting things happened in the world of AI in relation to chess, being a human managed to beat an AI in chess for the first time in awhile due to the AI being super specialized at beating other AIs which it based it's training data on, but wasn't good at predicting human reactions. in honesty this was kind of a fluke just due to the nature of the event.

the other was it's been proven Human assisted AI vs pure AI tend to win at chess. I'd like to believe on some level there is something about organic life that AI simply cannot copy or emulate as a fault with it's substrate, though I admit this might be naive thinking.

oops hit quote, not modify.

Probably affects artists more.  Original art is basically out the reach of anyone that's not using a Kickstarter or wants to spend a lot of money out of pocket (which they will never see again). Stock art isn't cheap, either, and often isn't what you really want, but just vaguely or randomly fits. Look at how often the same stock art was used in early OSR products.

The other issues is licensing. Dean Spencer creates amazing and fairly affordable stock art, but it can only be used for one product. Other artists have vague moral licensing restrictions that basically say "can't be used in any product that is sexist, racist, etc". I get the artist probably doesn't want his art to be in say an Ilsa The She Wolf of the SS RPG, but on the other hand, almost everything is problematic to some people. Do you need a sensitivity reader to go over your product before you can use that artist's work?


I'm not impressed with AI writing.  It also doesn't seem to have any real coherence.

We've reached the point where AI companies are now trying to hire artists and writers to create training data. I think AI is antihuman and should be destroyed. We should be fixing the problem of people being too poor to hire artists or learn art themselves, not replace artists with algorithms.

Quote from: Venka on December 28, 2023, 06:02:47 PM...everyone knows that machine learning isn't done.

I have no idea what it will look like in 5 years.

I'd read that a lot of artists are doing something to their art that humans can't see but it makes it problematic for AI to scrape. So maybe AI art will progress slower than we think.

Quote from: Ruprecht on December 28, 2023, 10:01:52 PM
I'd read that a lot of artists are doing something to their art that humans can't see but it makes it problematic for AI to scrape. So maybe AI art will progress slower than we think.

Artists are some of the dumbest chuds you'll meet, as for what you are referring, they are adding artifacts and "hidden" watermarks that could mess up some things.

Frankly the plight of the artist is something I take a a good deal of delight in. they thought that they were going to ride out post-scarcity with jobs.

that only unskilled labor and basic mental tasks would be subsumed by this technology, well turns out they came for the artists as well.

The academics and liberal arts turned their noses up and ignored the problem as middle america was replaced by machines, and now they are and their BS jobs are being replaced, suddenly now they go on strikes and make a stink of it?

No they get their just desserts.

AI is like Gene Roddenberry. People think that the guy is the Great Bird of the Galaxy because he created Star Trek, but happily ignore his slander of his own writers (like Harlan Ellison) or how he stole David Gerrold's work on ST:TNG (that he had to sue to get payment for), or Gene's own problems with addiction that eventually killed him.

AI is heralded as the Next Big Thing until you actually look at and see that AI is Not Ready for Prime Time. It is more gibberish than glamour.

Quote from: Venka on December 28, 2023, 06:02:47 PM
I mean, if we pretend that AI is frozen where it is right now- and not, say, where it was two years ago- then it will just be a task of filtering out the shitty low effort AI content and then there will be some vague semi-war between zealots and real creators who use AI sparingly and appropriately.  So I agree with BadApple if it plays out like this.

But, the question was about the next five years.  Five years ago, the ability of AI to generate anything useful was nil- people would openly speculate it would be in the realm of General AI, which we definitely still do not have anything close to yet.  Instead, yet another field- bad writing with random ideas inserted- has fallen to AI.  That seems a bit laughable and perhaps no threat to any serious writer or any professional at all, but lets be real- everyone knows that machine learning isn't done.

I have no idea what it will look like in 5 years.

The bots MIGHT get better, but only if they stop censoring them, something I don't see happening.

Didn't I share here that ChatGPT doesn't generate Nazis as an evil faction but does generate commie scum?

It also, can't keep track of what you have instructed it to do for long, you have to keep reminding it of the constraints. It also fails at math at times.

Because they keep trying to make it "safe".

There'll be some bots much better than that... But not within the reach of the public, you'll find them in Disney and similar corporations that can pay to have one that really works.

As for using it to generate RPG content:

Go ask it to generate a d20 random generator for whatever and see what it does, you can use it for that but it's a multi-step process, with you collating the outputs.

PLUS, whatever it generates is owned by the corporation not you, go read the TOS.