| 1 person marked this as a favorite. |
I just saw the last WOTC plans for our comrades on the other side of the d20 and one of their plans is to develop AI-GMs.
It got me wondering a few things:
Can it actually be a thing? A lot of adventures are quite streamlined and even if there's not tons of content to absorb (which is what AIs do the best) I'm wondering if an AI can actually be good enough to provide at will content to players when there's a dearth of game running. Maybe not for today, but for tomorrow...
How would you accept AI-GMs? Even considering that they won't be as good as human GMs (and I'm sorry to say that but I'm pretty sure an AI can be better than some human GMs), would you accept to play under AI-GMing? Will the lack of human interaction with the GM reduce the pleasure or will it be fine as long as you get along with the players?
Complementary GMing. That's one thing AIs do the best: helping us. A big part of our hobby can be fully automatized. Combats, for example, especially now that a lot of us are moving to VTTs. Just use an AI for combats and other simple scenes and, as a GM, you can use your time and energy on what's important (story, roleplay, character development). Is it something that would appeal to you?
In my opinion, it's only a matter of years before most of the media we consume (books, memes, films, etc.) will be mostly AI-generated. TTRPGs will follow suit, with AI taking a prominent role, just like VTTs have taken an important role since the pandemic.
That is not to say there will not be a place for human GMs. Just like people have continued to draw and paint even though we have cameras, and people still calculate things by hand even though we have calculators, people will keep GMing. But AI will be a huge support, and in many cases I think a human GM might not be needed at all in the future, except for those who want to have that old-school feeling, just like some people right now do not want to use digital dice-rolling tools because they like rolling physical dice.
Already people are generating art and text with AI to support their storytelling. As long as we view AI as a tool this is the way to go in my opinion.
| 2 people marked this as a favorite. |
with ChatGPT is it seems technically possible now. They would just have the spend the money to put it all together. It may take a few years and maybe it is not economic yet. But it is not impossible anymore.
Would you pay money with a few of your friends to take on a Computer GM?
I would.
When they‘ve passed the Turing Test, why not?
| 10 people marked this as a favorite. |
Called it months ago. It's the logical consequence after WotC showed 0 support for human GMs in years - they need a game with a GM, but being a GM in 5e sucks, so let's automate them.
Whether they're any good, we'll see.
| 2 people marked this as a favorite. |
Would you pay money with a few of your friends to take on a Computer GM?
I would.
I wouldn’t. I wouldn’t pay to play, and I don’t “take on” the GM, computer or otherwise. I get it, you would. But I think it sets up just what WotC are looking for.
| 1 person marked this as a favorite. |
In my opinion, it's only a matter of years before most of the media we consume (books, memes, films, etc.) will be mostly AI-generated. TTRPGs will follow suit, with AI taking a prominent role, just like VTTs have taken an important role since the pandemic.
That is not to say there will not be a place for human GMs. Just like people have continued to draw and paint even though we have cameras, and people still calculate things by hand even though we have calculators, people will keep GMing. But AI will be a huge support, and in many cases I think a human GM might not be needed at all in the future, except for those who want to have that old-school feeling, just like some people right now do not want to use digital dice-rolling tools because they like rolling physical dice.
I think you are drawing very different conclusions to me about the difference between a tool, and AI given your use of comparisons like drawing/painting and cameras.
Already people are generating art and text with AI to support their storytelling. As long as we view AI as a tool this is the way to go in my opinion.
I would not agree that generating “art” and text “supports” storytelling. As long as this is how you and others view AI, then it is the way you will most certainly go.
I don't think there's a point in being against AI-GMs, as they will be a thing whatever we think.
I'm wondering how we can "embrace" them. I mean, companies like WOTC are full on them for monetary reasons, not for actual improvements of the hobby. If we don't react fast enough, they will certainly have their way because of the lack of serious competition. But if we can create our own vision of AI-GMing, one that is taking the best of human GMing and AI-GMing, we could actually improve our hobby.
In my opinion, complementary GMing is a boon for paid GMs. Currently, they struggle to make a living. But if you use AIs for complementary GMing, you can handle more tables simultaneously, improving significantly your earnings without asking for more money.
I'm convinced that paid GMing is part of the future of the hobby. It's actually what WOTC is trying to do: Earn money from players. But at least with paid GMing, we have an improvement of the quality of our hobby (I consider a professional more committed and experienced than an amateur, which is not always true but tends to be).
|
| 16 people marked this as a favorite. |
I think people often make mistaken assumption here that people only GM out of necessity rather than it being fun :p
But yeah, thing is that AI GM wouldn't be able to develop NPC PC relationships, have inside jokes with players, laugh with players at their antics, etc. At most it would be able to generate storylines, bit repetitive dialog and run the combat math, just look at stuff like AI Dungeon for reference. And then there is question of "can AI make fair rulings for table or make sure everyone is having good time"?
(writing AIs are able to generate really good looking short stories form samples, but the longer novels need heavy curating or you will get lot of droning. I think AI tech hypists kinda ignore lot of bad things about current AI progress with assumption that future AI will solve those problems inevitably
That and main thing that AI will actually replace is programming, because while art and writing comes with weird bugs and questions about "well does ai generated art have worth or not", programming is something that AI can already do extremely well depressingly enough for anyone who wants to be a programmer
But yeah thanks wizards for keeping this conversation alive, I'm kinda sick of this topic in most discords I've been at :'D)
| 1 person marked this as a favorite. |
I don't think there's a point in being against AI-GMs, as they will be a thing whatever we think.
I think it’s pretty clear they are a thing, and I’m still not into them in the slightest.
I'm wondering how we can "embrace" them. I mean, companies like WOTC are full on them for monetary reasons, not for actual improvements of the hobby. If we don't react fast enough, they will certainly have their way because of the lack of serious competition. But if we can create our own vision of AI-GMing, one that is taking the best of human GMing and AI-GMing, we could actually improve our hobby.
And then they created their own weapon to outweaponerise the evil Wizard’s unthinking thinking weapon, and they rejoiced for the two point minus six negaseconds it took them to realise this was a very, very bad idea….
| 10 people marked this as a favorite. |
In my opinion, complementary GMing is a boon for paid GMs. Currently, they struggle to make a living. But if you use AIs for complementary GMing, you can handle more tables simultaneously, improving significantly your earnings without asking for more money.
I'm convinced that paid GMing is part of the future of the hobby. It's actually what WOTC is trying to do: Earn money from players. But at least with paid GMing, we have an improvement of the quality of our hobby (I consider a professional more committed and experienced than an amateur, which is not always true but tends to be).
Hmm. I’m convinced the future is spoiled with a move to paid GMs. I know they are a thing. I’ve seen them touting on their websites.
I’m going to just post this here because I can’t let it go, as it is derailing the thread. Ravingdork was in a three or more years ago I think about paid GMs, and I’m pretty sure I ranted there, so maybe we can dredge that up, but here goes:
Bringing capital into the “game shared by friends” changes the social reality into “friends and someone who is paid”. Right off the bat, there is a necessary interrupt in the social contract. One of these people now has a concrete monetary value placed on their agency, their creativity and their ideas that the others do not. In fact, given the economic framework under which most of us live, that one person is now valued higher than the rest.
That value, once entertained cannot easily be removed.
Player “Oh, we can’t pay you this week.”
Paid GM Oh..uh…no..that’s ok. We’ll play next time, you know, when you folx can pay” OR “I guess I could do a freebie” (you know like I could anyway, but weirdly, sometimes I’m of value, and sometimes I’m…not!?!)
Player “Well it’s not that we don’t value your..ya know…ideas it’s just that…” (fumbles d20 nervously)…
It’s super gross that one person’s ideas and creativity are given primacy “because they are running the game, taking the time to blah blah blah…”sure, there’s the prep and the responsibility, but what about the creativity and ideas of the players? Why would you pay someone to share your ideas and…hang on…how is that sharing?
And as for responsibility - in a “friends hanging out” situation the responsibility is equally held - to be friendly as a player, to be courteous, respectful, not to make the game all about you. - like in most social situations devoid of employee/employer/contractor relationships. What next? Or is it already a thing? Paid players vying for the next big actual play? Extolling the virtues of their virtuosity? Chosen for their chops and paid accordingly? How far down the ladder does this go?
Which brings me to the next point. Adding an economic tier necessarily creates haves and have nots. Can’t afford to play with your friends and their megabucks GM? Too bad. But wait, that’s real life Oceanshieldwolf, and the market decides, wallets have power and if Wizards are going to do it then let’s beat them at their own game!!! This hobby had the potential to be amazingly therapeutic, transformative and evolutionary, but capitalists keep telling you that professionalism will improve the hobby, and not to worry about the masses of people who don’t already play because their socio-economic status makes it improbable that they’ll ever interact with it, share their stories inside it or help it to become way more that it already is. I don’t think I need the hobby improved if I need to pay for it to get better when it has made amazing strides in 50 years.
And yes, paid GMs can do charity gigs, or take on players with hardship, but really, all that just glosses over the sadness that adding money to anything always does. It breaks intimacy, obfuscates personal relationships and generates confusion about social motivations. Some people will disagree and say that the opposite is true, that money clarifies positions, relationships and motivations. All I can say to them is I completely and wholeheartedly disagree.
I just had a thought. So hang on, what about paying someone to draw your character? That’s ok, because that’s a service…right? And if you can’t draw, and nobody you know can draw or will draw your character for free, you have to pay someone…right? Well, yes and no. You don’t have to. But you can. And I understand that. There are trained and innate skills involved. But hang on OSW - can’t a GM have innate and trained skills at GMing that they could and perhaps should be paid for? Aren’t their people who have no time to run a game, groups without one person who will run the game and they just want to freaking pay someone to run the game for them?!?
I guess. But they probably shouldn’t. For all the reasons I outlined above. Have fun gaming, and remember there’s no wrong way to play RPGs, unless one of you is being paid. ;)
GMing isn’t a “service”. It’s a privilege, and responsibility and a way of life. It’s not a job. Once it’s a job you’ve lost the privilege, and your responsibility is bought and paid for. And your way of life has become a capitalised trap.
I think you are drawing very different conclusions to me about the difference between a tool, and AI given your use of comparisons like drawing/painting and cameras.
What's your conclusion? Both photography and painting are art forms, wouldn't both human-GMs and AI-GMs be valid TTRPG forms?
I would not agree that generating “art” and text “supports” storytelling.
Why not? And what do you think of virtual dice rollers and fantasy name generators that have been around for much longer?
| 1 person marked this as a favorite. |
I think people often make mistaken assumption here that people only GM out of necessity rather than it being fun :p
But yeah, thing is that AI GM wouldn't be able to develop NPC PC relationships, have inside jokes with players, laugh with players at their antics,etc. At most it would be able to generate storylines, bit repetitive dialog and run the combat math, just look at stuff like AI Dungeon for reference. And then there is question of "can AI make fair rulings for table or make sure everyone is having good time"?
[Emphasis mine]I guess you said it way better than I ever could. Especially that second one - because the game is a shared thing where we are having fun together. I mean, I guess if the AI is having fun, and everyone else is too, then it’s all thumbs up?
| 1 person marked this as a favorite. |
I think you are drawing very different conclusions to me about the difference between a tool, and AI given your use of comparisons like drawing/painting and cameras.What's your conclusion? Both photography and painting are art forms, wouldn't both human-GMs and AI-GMs be valid TTRPG forms?
I just don’t think the comparison is valid.
I would not agree that generating “art” and text “supports” storytelling.Why not? And what do you think of virtual dice rollers and fantasy name generators that have been around for much longer?
Because the I think the term “supports” is misapplied. It’s not supporting. It isn’t providing anything that I believe has any worth to the storyteller’s approach to telling their story. I guess we can just disagree here.
As for virtual dicerollers: sure, I have no problem with a digital RNG. I’d prefer a dice if in person, but for PbP I “rely” on them.
Fantasy name generators are an anathema. If you can’t name your own character, why are you playing? Given the choice between certain death and naming your pet, everyone names their pet every time. It isn’t that hard.
I think the AI GM will be there in the future but it will create a new niche like VTTs done.
Using VTTs for example. They don't substitute the live in place gameplays but they allowed to do a remote gameplay way better. Sometimes you can use tablets/notebooks to play face to face using VTTs to help but works like calculators, they are there to help but aren't necessary are just one more tool to help GM and players only. So in the end the VTTs created a new niche that don't substitute in person gameplays but allowed to those who don't have much time or are too far way to play.
The same could be said to AI GM, probably the felling would be completely different from a real person GM, less jokes, less humanity, less cry for GM (kkk). This will solve the situation where no one want to be a GM but this is a niche, tables with real GM will continue and will be preferred by many other people maybe they can use AI to help but not to substitute in these cases.
| 1 person marked this as a favorite. |
AI will be able to scratch an itch, much like RPG video games do now. But look at "simple" first person shooters like Call of Duty: most of those campaigns have diverse, dynamic combat settings and many interactive NPCs, yet even with one PC & a horde of creators the story needs to be railroaded on a mostly linear map with minimal RPing options if any at all with all the goals prescribed sans player input.
It'll be a while before AI can replicate anything so sophisticated, much less a multi-PC scenario (much less campaign) with NPC dialogue & drama with player-driven goals and diversions.
That still leaves room for randomized dungeon crawls and pure combat, much like Diablo with more advanced mechanics & maybe even original art and monster reskinning. "Is that a Shambling Cloud Giant necromancer?!"
That could be cool to pass some time, but that's not the TTRPG experience for me, more like MMORPG stuff, though with the benefit of being replenished with fresh content via the AI.
As mentioned above, a hybrid version seems more effective, inevitable, and to me quite welcome. Who knows how much of the labor an AI could shoulder for you, generating aids, doing research, running numbers, and controlling VTT combat/lighting/time management. That could go truly outstanding places where maps and encounters could be insta-generated to match the GM's input if the players go on a tangent, yet with GM veto & editing to fine tune it. The GM could program in the personalities and phrases of NPCs, taking over for the more important ones or when the AI needs guidance (and likely often surprised by creative AI moments).
An AI could heighten a GM's abilities in most every facet, and much of that swiftly too. Not that I'll necessarily have the funds or lifespan to see it. :/
| 9 people marked this as a favorite. |
Personally, I actually like GMing, so I wouldn't use an AI DM; if my group lacks a DM, I'll happily fill that role
However, until capitalism stops being a hellhole that monetizes literally every aspect of existing, I'll be against them for the same reason I'm against AI generated images and stories being sold for commercial gain
| 2 people marked this as a favorite. |
Video games have been doing this for years.
The campaign won't be quite as much fun because the AIGM won't really be able to improvise well. Much like a video game the plot will be pre-scripted. But there are human GMs that have to fall back on that too.
I also think it would be best if the players were able to override the GM in some fashion. Because there is always the risk that the players want to go in a different direction than the AIGM would allow.
| 1 person marked this as a favorite. |
If they can do AI-GMs well, I don't really have a problem with the idea. The execution is critical, otherwise it becomes essentially just a theater of the mind CRPG.
The holy grail would be an AI that would allow a team of top-tier roleplayers like Critical Role to dive into a world with all the creative problem solving, roleplay, and compelling story arcs that you see from the best GMs. I am HIGHLY skeptical that they'll be able to do anything remotely approaching that within the next decade or two. Until it reaches that point, it will be a novelty game mode for when the GM needs a break, or for groups who just want beer and pretzels fun.
My concept for a procedurally generated dwarf mine game mode would work pretty well for such an implementation.
In any case, it's far better than AI-art-theft on an industrial scale.
|
I'm pretty sure that, at least for a long while, AI-GMs will work far, far better for very precise rule systems in very bounded situations.
So I think they'll work far better for One D&D doing dungeon crawls than they will for, say, running a War For The Crown campaign using PF1 or PF2 rules.
But for that (possibly very large) subset of the community that really just wants to go into the Dungeon and beat up on the monsters they'll probably work quite well fairly soon. Well enough. As others have said, better than some human GMs.
So I kinda expect the community to split. On the one hand will be groups running a reasonably simple system and doing dungeon crawls with AI-GMs on line. Over the years the AI's will get better and better and what they can handle at all and handle effectively will grow.
On the other hand we'll have more or less what we have today. A whole huge mismash of games, campaigns, gaming styles with humans as GMs, aided by technology in some cases and going old style in other cases. Some of it online using VTTS and maybe AI-GM-helpers, some going old style and sitting around a table together. Quite likely this community will NOT be playing D&D from WOTC any longer, I expect WOTC will make sure of that.
The 2 communities will be largely distinct, just like MMO/TTRPG/Video game communities today are largely distinct (with lots of overlap, obviously, but still largely distinct).
I also imagine that the new AI-GM community will end up both larger and spending a lot more money than the left over TTRPG community. Just like more people play MMOs than play TTRPGs.
But hopefully the TTRPG community will still thrive and survive, even if it no longer plays D&D from WOTC. It may be bigger than it is now as it draws in players from the AI-GM world, it may be smaller as it loses players to that world, but it will quite likely outlive me (but I'm an old fart, so it may not outlive YOU :-()
You can already have an AI run a freeform game, but you can't have an AI do it well. Even just asking an AI to spit out campaign ideas, they're pretty painfully generic unless you give it the creative part. The AIs that can keep track of things repeat themselves insistently, and the "creative" ones wander off and lack narrative or thematic coherence.
The thing is, giving the AI some "concept" of rules is really difficult. It doesn't understand what it's writing, so what do you do... have a second AI that looks at the text to decipher where a check or saving throw is appropriate? I'm curious what their training data for that is going to be. How do you get data on "good GMing"? On handling multiple characters as AI-run, and multiple characters as player-run? Without some good training data, I'm expecting it to take a lot of work getting it to work well. I think we're in the land of dancing bears and overly optimistic tech ambitions.
Maybe the terms of the VTT will let them harvest your games for training data, though. Maybe they train it on adventure modules, which might have standardized enough formatting for what checks are expected to be picked out from AI-generated text.
We'll see, though. Getting a group is hard, so maybe this will catch on even if it's not very good.
EDIT: That's a good point, I definitely expect AI to be able to run a perfectly functional generic or even themed dungeon crawl. You can do that in the other order- use a regular algorithm to generate the monsters you have in the room, pick combat actions from a set of options, and then just have the AI generate the text to describe what's happening. Given they said "stripped down", I think that makes a lot of sense- combat mode only, no overly open-ended spells. The AI can just be a replacement "graphics engine" for a video game.
| 2 people marked this as a favorite. |
I studied machine learning rather than artificial intelligence, but a lot of so-called AI is really machine learning where an algorithm trains on a data set. If Wizards of the Coast creates an online roleplaying system where the character takes instructions, such as attack this opponent, move to next opponent, or cast a particular spell, rather than just makes dice rolls, then the data set could be all the actions of a particular kind of enemy character, such as all 1st-level orc warriors, run by a real DM, and the computer could develop a decision tree to mimic that behavior. A little more sophistication would let the enemy use different tactics against a fighter than against a wizard, but the computer might have to metagame by reading their character sheet in order to tell the difference between a fighter and a wizard.
Bluffs and diplomacy would be harder to model, but WotC could set it up like a computer game with pre-scripted dialogue, so the bluff and diplomacy rolls would just be dice rolls on the script.
But then an automated D&D game would merely be a token-based video game with D&D classes and abilities. The fun part of improvising tactics and planning in advance would be gone.
I think people often make mistaken assumption here that people only GM out of necessity rather than it being fun :p
But yeah, thing is that AI GM wouldn't be able to develop NPC PC relationships, have inside jokes with players, laugh with players at their antics, etc. At most it would be able to generate storylines, bit repetitive dialog and run the combat math, just look at stuff like AI Dungeon for reference. And then there is question of "can AI make fair rulings for table or make sure everyone is having good time"?
(writing AIs are able to generate really good looking short stories form samples, but the longer novels need heavy curating or you will get lot of droning. I think AI tech hypists kinda ignore lot of bad things about current AI progress with assumption that future AI will solve those problems inevitably ...
I GM for fun myself. I run Paizo modules, but have to modify them and improvise on the spot because my players are very creative. I enjoy the challenge. And my players defy expectations.
For example, in my Ironfang Invasion campaign, the Ironfang Legion led by human-hating hobgoblin General Azaersi is confused by the heroes, none of whom are human. The party is sympathetic to the Ironfang cause of wanting a nation of their own where they are not second-class citizens to the humans, but they cannot tolerate the hobgoblins enslaving humans in the lands they conquer. For example, in Prisoners of the Blight they encountered four Hobgoblin Commandos mind-controlled by blight. They fooled one, put the second to sleep, knocked the third unconscious, and stabilzed the fourth after lethal damage to cure them of blight rather than kill them. Some Diplomacy formed a temporary alliance against the blight, aided by the high-rank commandos knowing the honorable repuation of the party.
I was delighted how my players develop a unique narrative. Yet I doubt an AI GM could handle the nuance that let them temporarily ally with their primary enemy.
On the other hand, setting up an AI to assist a human GM could work well. The GM could create a war band of orc warriors, run the players scouting the war band and setting up an ambush by himself, but when the situation comes down to combat, he could ask the AI to run the combat tactics of the orc warriors using a built-in decision tree, with edit options (Yes, charging is a good choice, but charge the wizard rather than the fighter) to correct the AI's blind spots.
|
| 1 person marked this as a favorite. |
This thread surely won't end up toxic... I hope.
So many of the conversations about AI advancements I've seen of late go south SUPER fast, especially when self-described full-stack programmers get involved, act defensive, and try to act like they're some authority on it, so let's please be mindful to respect each other's opinions and viewpoints here.
This thread surely won't end up toxic... I hope.
So many of the conversations about AI advancements I've seen of late go south SUPER fast, especially when self-described full-stack programmers get involved, act defensive, and try to act like they're some authority on it, so let's please be mindful to respect each other's opinions and viewpoints here.
Hey now. I've already joined the thread. so...
-----
Anyway, this somewhat reminds me of playing the Pathfinder Action Card game. The game box and the rulebook is the GM. All the people at the table are players. It very much plays out like a dungeon crawl. Heavy on action, light on roleplay.
That is very much the type of thing that an algorithm is good at.
| 2 people marked this as a favorite. |
I think this is the perfect plan for a Skynet. Dominate/eliminate the nerds, without them it's much easier for humanity to fall... (hell why didn't I think of that before?!?!?), seriously now, like everything that can be computerized, it was just a matter of time ,but I believe that only our children or grandchildren will be able to enjoy an AI GM "as" good as a human being. If I had it today at a fully functional level, and it wasn't a fortune to have, I would use it without any problems...
Just an addendum, I would use it if I didn't have the option of a human GM obviously, I've spent years just looking at books I bought and never played, some on my shelf know only, the shelf. Personally, it's not what I would be completely satisfied with, I like having my friends on my side, so I don't even like playing online, even though I'm playing Agents of Edgewatch with sensational people, it's not the same thing. Nor would I pay for some GM. It would be a last (and desperate) option. A certain Stan once said: "Seeing breasts on a screen is good, but nothing compares to feeling them in your hands..."
While the RPing hurdle has been addressed here (and IMO seems insurmountable beyond scroll down menu options), another facet is tactical acumen. Sure, an AI will likely make balanced encounters like a proficient GM, maybe even maps that sync well with the monsters' abilities, but there's such an enormous amount of combat options in TTRPG. Even in modern scripted video games w/ far fewer, terrain and AIs can often be exploited.
Think about an AI having to adjust for wall spells, blindness, illusions, and so much more, not only on the fly, but in how they design the terrain (including proximity of other encounters, surrounding terrain). Heck, just kiting has proven difficult for AIs to overcome. And assuming creators continue making content, more and more permutations become possible. Some the AI itself might exploit if it's too much of a rules lawyer which doesn't it kinda have to be? Doh.
| 2 people marked this as a favorite. |
To me a game without a GM is a board game not a TTRPG. There can be board games that mimic the TTRPG experience (Gloomhaven was based on Pathfinder 1e, after all). Hasbro obviously doesn't have a problem publishing board games (nor should they) but without a human referee's ability to adjust on the fly, ad lib, and "yes, and" their players you just don't have the same thing.
For a lot of people I suspect "we play through the module with a computer GMing" is going to be fine, but I don't know a TTRPG player whose best experiences at the table weren't brought on by something going hilariously off the rails and all the human participants making the best of a chaotic situation.
The long and short of it is that improvising with a computer just doesn't seem fun.
| 14 people marked this as a favorite. |
My prediction: When you can get a human GM, only the GM needs to buy the books. When you need an AI GM, everyone's gotta subscribe.
|
I think an AI run game is going to feel a lot like either a "choose your adventure" or like a video game. Compensate the shallowness of plot and interaction with nice visuals.
I also think that eventually AI will get better at this, but it's far more likely that the breathtaking AI that'll all get us hooked was developed by some firm we'd previously never heard of, using an innovative approach that a big slow corp like Hasbro didn't really think of trying.
My prediction: When you can get a human GM, only the GM needs to buy the books. When you need an AI GM, everyone's gotta subscribe.
At that stage it'd seem more like Baldur's Gate multiplayer extension patches with AIs implementing some of the development. That would be interesting, but as many have noted, not in a TTRPG way. It could lead to mass development, especially as communities shared the better creations, even down to specific monsters, art, and moments.
That said, my dream had been for Neverwinter Nights to be DIY Baldur's Gate so I was sorely disappointed. I have to wonder what a modern or near-future AI could do with that engine and similar artwork, etc. That'd be welcome, albeit not TTRPG level. Of course modern gaming standards have bypassed that, both in tech & narrative, but I think the lure of endless encounters would appeal to many, if only in a hack n' slash way. People still play old Diablo games, and modern AI can randomize better encounters than those.
I get that Hasbro is motivated by one of the fundamental problems of this game is that it can be hard to get a group together which is compounded by situations where nobody wants to GM. But "Nobody wants to GM" is basically the driving principle of modern board game design.
I'm sort of sad about the idea of a world where people don't feel driven to want to learn to GM out of the love of the hobby itself.
| 8 people marked this as a favorite. |
....
I guess. But they probably shouldn’t. For all the reasons I outlined above. Have fun gaming, and remember there’s no wrong way to play RPGs, unless one of you is being paid. ;)
Most of your post is political.
The public really need to understand that capitalism is good, if and only if, there are alternatives and other options for people. Then the market can function and ensure that people are getting what the people what.
If a market is being dominated and other things excluded then its is bad. Paid GMing and AI GMing is good because you will always have other alternatives. You can just use your books and keep going. But yes when one company attempts to use their market power to dominate and abuse their competitors it is time to put a stop to things.
GMing isn’t a “service”. It’s a privilege, and responsibility and a way of life. It’s not a job. Once it’s a job you’ve lost the privilege, and your responsibility is bought and paid for. And your way of life has become a capitalised trap
What gaming is not is a morality that you can impose on GMs. We all do it for slightly different reasons. I'm very happy for people in this industry to earn a living.
| 1 person marked this as a favorite. |
If AI-GMing can get beyond what Wrath of the Righteous can deliver it will have value. If you can interact with it using natural speech and a VTT it could easily be the first way people experience TTRPGs.
As somebody who mostly GMs for friends, I would probably enjoy blowing off steam by running 4 min-maxed characters in an AI-run game.
|
As an Economist, I wonder why WotC thinks they have a competitive advantage. I imagine AI-DM will be closer to a video game than a TTRPG. It sounds like they are planning it in a combat-only role, and I'd rather have Diablo or WoW mechanics than any version of D&D. The advantage of D&D is doing weird stuff. I have a professional interest in thier plan to compete with video games.
My prediction: When you can get a human GM, only the GM needs to buy the books. When you need an AI GM, everyone's gotta subscribe.
This, Plus Grinding! We'll have people who play all day, every day, just like in WoW. They'll have maxed out level 20 characters, running quests over and over again for the best gear. Then they'll join human-DM games with super-powerful characters.
Letting players transfer characters between human and AI games is the only advantage WotC has and it's predatory as hell.
I feel like "not allowing GMs to disallow characters" whether it's because they're suspicious about the provenance thereof (a la side-eying rolled stats) or they don't fit the campaign or they seem likely to cause intraparty conflict or whatever is a dealbreaker.
| 8 people marked this as a favorite. |
I guess the comedy upside of this is that since AI programs like ChatGPT have the ability to be very confidently wrong about math, a bunch of people are going to have to argue with a computer about whether a 15 hits or not.
I guess the comedy upside of this is that since AI programs like ChatGPT have the ability to be very confidently wrong about math, a bunch of people are going to have to argue with a computer about whether a 15 hits or not.
LOL indeed.
That is a feature of algorithms. When an algorithm comes to a result, it is by definition the correct result. If the result is not accurate to reality and what the humans want, then the algorithm is not the right algorithm for the job rather than that the algorithm was wrong.
This is why the players of an AI-GM game need the ability to override the GM when necessary.
Contributor
|
| 2 people marked this as a favorite. |
This whole discussion gets into a number of interesting sub-topics:
1) How do you incentivize people to GM?
There are some of us who are called to it, but by and large, the Organized Play program has shown that getting people to GM at Conventions and elsewhere requires incentives. I do believe that the Achievement Point system that they use does a good job getting people to GM.
But even with that incentive, GM training and recruitment is something that we constantly do at my venue, because we have 2 regular GMs and three tables to cover.
2) How do you balance those incentives? What changes when you 'pay' people to GM?
There are professional GMs out there who are paid money to GM for others. This was something that I fantasized about doing myself until someone actually offered me a significant amount of money (significant to me, anyway) to GM for their group, and I realized that I didn't really want to become an employee of my players. I turned them down, which was a difficult decision as at the time I could have really used the cash.
If you make GM incentives too high, you can get people GMing for the wrong reasons.
3) What do you want from a GM?
Are you looking for a video game with a few skill challenges and a lot of combat? An AI could handle that fine? Are you looking for in-depth interactive storytelling and laughing with your friends? An AI GM might find it harder to scratch that itch.
Hmm
|
|
| 4 people marked this as a favorite. |
I feel like "not allowing GMs to disallow characters" whether it's because they're suspicious about the provenance thereof (a la side-eying rolled stats) or they don't fit the campaign or they seem likely to cause intraparty conflict or whatever is a dealbreaker.
Hasbro sees DMs as employees who pay them.
| 2 people marked this as a favorite. |
There are a lot of important ethical questions about the way we implement AI, but the technology itself is here and it's getting better at a fast pace.
I think the potential for AI is amazing for roleplaying games. Getting a good GM is hard and a huge obstacle to playing TTRPGs. We have a lot of people interested in playing but no one interested in GMing. It's one of the reasons I've still only played a handful of sessions of TTRPGs my whole life.
The ability to just have a DM, at will, would be huge. I might finally be able to play TTRPGs.
|
|
| 2 people marked this as a favorite. |
Fantasy name generators are an anathema. If you can’t name your own character, why are you playing? Given the choice between certain death and naming your pet, everyone names their pet every time. It isn’t that hard.
I'm horrible at naming things. I have a large pet cemetery to prove it. All the tombstones say things like here rest ___ or X we miss you.
But when I do use name generators, I never use the exact output.|
|
| 1 person marked this as a favorite. |
The whole reason I got back into TTRPGs was to get away from computers. I have no interest in using VTTs even though I programmed one in second life. Even less interest in an AI GM, specally a pay to play AI GM.
If you like the idea of playing against an AI GM to feel like your beating the machine playing cyberpunk seams more approaite.
|
Contributor
|
| 2 people marked this as a favorite. |
Oof. Just getting caught up in the tangents about paid GMing. A lot of hard work can go into this hobby. It's natural that some would want to monetize it. Especially for strangers and new groups. The paid games on the official pathfinder discord are pretty popular it seems. I mean, I wouldn't want to pay for games but I'm sure those involved are having plenty of worthwhile fun.
No doubt. And it was something that I was dead certain that I could really well. My concern was about how the dynamic would change me, and how it would affect the relationships with my players. I decided at the time that I didn't want to experiment with it.
But considering how much time I put into each game, and the joy that I give... While I didn't want to turn my hobby into my job, I don't begrudge people who decide to monetize their talent. I just know that money changes some of the expectations, and can really complicate a social interaction.
Then again, maybe I'll change my mind later and make GMing a retirement gig.
| 12 people marked this as a favorite. |
I don't know about full paid GMing, but I think we should normalize buying Doordash for our GMs now and again.