Personally I only really see AI as useful for image generation here. There's already a ton of name generators and tables to roll on for little character quirks for an NPC and I simply will never trust the current generative stuff to string together a useful narrative or dialogue.

Quote from wagnarokkr >>

Regardless of what the numbers actually are, responding to someone that says “Thing A is bad” by saying “Unrelated Thing B is just as bad” is a very silly argument.

 While the specific example is bad because they're not comparable, people's moral stances vary wildly from person to person.  

Personally, I dislike the power consumption complaint, because it's a lack of power overall that's at issue.  Like telling people to turn off the lights when they leave the room, while using super low energy LED lights that use barely any power.  Not using AI won't solve our power problems while people continue to rail against renewables that can be rolled out cheaply and quickly for example, but it's often the same people complaining about both things. The legal framework for plagarism definitely needs to be fixed when it comes to AI content, 100%, but the energy cost wouldn't be an issue if people stopped railing against things like reopening nuclear power, or deploying solar farms in the desert either.  Our world runs on electricity now, including the cars and busses, and we need to rebuild for that reality and stop trying to "get the most out of old assets" like coal fired power.

Regarding D&D and AI, the argument for "Don't use it, it's evil" is the same problem in debate that people have with many new technologies.  A technology that's been shown to make people's lives faster and more comfortable, will never be abandoned because people think it's damaging.  Plastic is a perfect example. It's destroying the planet, yet use is accellerating even today.

AI being used by DMs is inevitable.  People can hate that reality, but it is where the world is going as a whole, and D&D isn't a special niche who can avoid it.  Like writing software however, it's insufficient to do it alone.  You will need to treat it like a random generator that makes it's own random tables on the fly.  It can be useful, but it's not a 90% fix for your own work.  It's probably not even a 40% fix.  It can spit out a bunch of 'ideas' many of which are nonsense.  

You need to know your table.  You need to know your world.  AI can steal info on the existing worlds from things like the Forgotten Realms wiki, but so can you.  All AI is really useful for at this point is as an English major.  They can write prose, but they have no idea what it means or if it makes sense.  That's completely on you as a user.

Quote from GnollitAll78450 >>

Your friend was right, you weren't learning. You were also burning through a small neighborhood use of power to not learn. 
I strongly encourage you to be upfront with your group saying "Hey, i am still learning, have patience while i puzzle this process out" and if they are a good group they will be understanding.
Gen-Ai is not like the random name generators that have been around since the early days of the net, they use WAY more resources, and are trying to make you reliant on them, and not just 'coming up with an elven sounding name for an NPC who will be used once.'

I am going to expand a little, about specifically, Not learning, and how it isn't just an A.i. thing.

If you just found a bunch of encounters online and then plugged them into your adventure, but didn't take time to look at and puzzle out why it works, that is also a form of not learning, but it is much easier to start learning from. 
You need to start asking the question of, "why does this work?" and/or "Why didn't this work?" because despite being a math heavy game D&D is more than math. 
In old D&D, there was a list of random encounter tables that made absolutely no sense, and were just there to be there. You could get 4 litches showing up when a barbrawl happens and unless you are ready to do some really creative story surrounding why 4 litches were in a tavern, it was just absurd. it was a "Why didn't this work?" moment. A little thought of 'there is no reason for them to be in this tavern'  is the answer to why it didn't work. You answer that, and then you start asking 'what would work?' and if it was level appropriate, but thematically inchoehrent then you know what to change. just change "litch" to "wizards" and maybe take off some of the undead specific stuff and replace it a little, then BAM. 4 very powerful wizards just happened to be in a tavern. 

I will admit, that is an extremely simple example and the answer to 'why didn't this work' can be very complex, but it gives you an idea of how to start learning. Yes the math can be hard, yes the vibes and balance can be strange and inconsistent. But technology need not be absent from it. There are CR calculators that are not GenAI based and just a simple low power algorithm that can help with the math. if you have the formulas from the DMG, and you just use a calculator, then you are still learning the process. 

I have a math disability. I know how much numbers can be a pain, but certain physics concepts that are very math heavy, i understand, even if i can't calculate them in my head. Here is the fun part. Most physists can't either. They use technology, just not the technology you are asking about. 
There are other technologies. There are other resources, there are other ways to learn the process without melting your mind.

You said you started relying on it to make encounters that fit your campaign OP, what aspects did you need help fitting better? If you tell us, we can help you find ways to do it yourself, ways for you to learn. and when you learn the rules, you then can learn when it is best to break them to wow your players. 

Circling back to the Litches in a bar, you break that 'rule' and people ask 'why are they there?' and you can respond ' I am sure your character is just as confused. Perhaps they should find out?' and that can a hook for "The red Wizards of Thay are sending in spies to the area. They might be looking for something. We don't know what, but it is best they don't have it." 

Edit: Rereading this, some people might think i am deflecting from the Gen-Ai thing, so i will clarify my stance. If said the words that accurately expressed my hatred for Gen-AI i would likely be incarcerated. My hatred for the thieving little plagiarism boxes that are the current Gen-ai models is intense, special, and not for polite company. It is a technology designed to steal from, then starve creatives.
 I think it is, without hyperbole, made with malice against the every creative aspect of the human soul.

New research out of MIT shows we incur a 'cognitive debt' that is nonrecoverable when we have ChatGPT write things for us. Even when we simply use it to provide us with answers instead of going through the processes of actively searching for things.

The research used advanced neurological imaging to show that we are not using and exercising certain parts of brains when we use it that we would otherwise be using and exercising. The same research also found use of it erodes users' capacity for critical thinking. (Probably why we see so many of its proponents using it to generate incoherent or incorrect content and their then being incapable of seeing these problems.)

You can't argue with the science no matter how much those who will most benefit financially from its being used like to act as if it is going to save us. Telling us what are nothing more than the sort of 'prophecies' cults trade in when they tell us no matter how much it's harming the environment now... it's totally going to help us save the planet!

DM-ing is a creative endeavor. When we are small children we are perfectly capable of coming up with games of our own invention. I ran my first D&D game when I was still a child. I got into the hobby before I had even finished primary school. I used the books on my shelves and the illustrations within them as well as my favorite movies at the time to inspire me.

As for names for NPCs, I have been using the Albanian words or phrases for concepts as NPC names for years. Or, just rolling names up on one of the many random tables out there in one of the many great zines or on one the many great blogs others in the hobby have made available for us. The best of these are made up of names worthy of characters even the best authors of the genre might dream up. Why would I need a machine to make something up for me? Assuming that machine isn't just stealing what it gives me?

I will add onto the "not learning" thing.

A large percentage of groups these days, run only the provided campaigns.  They don't write their own stories, and they don't want to.  They play the published campaigns, then complain when WotC doesn't provide them with new campaigns.  These are not DMs that are 'learning' beyond their first campaign.  They're reducing their workload by using a pre-built campaign, and just playing it as it's written.

Those people using AI to write their own campaigns are actually more creative than people just following a provided campaign.

It's the big difference I've seen since Basic/2nd Ed/3rd Ed, and the modern 5E game. (I didn't play 4E.)  The over-reliance on someone else to provide the story.  Back in the TSR/Early WOTC day, you'd use provide adventures in your game, throwing them in where it was appropriate and adapting them to fit.  There were campaign settings, not campaigns.  The publishers provided the world to reduce worldbuilding, so you could focus on writing the games.  These days, the worldbuilding from WotC is getting thinner and thinner, like removing the lore from the Monster Manual, because they want to focus on providing people with the next Curse of Strahd best-seller.

When you bought the Menzoberran box set, it had two books of world, and an adventure to get you started.  About the only "full campaign" I can remember was the Rod of Seven Parts, and that was their (somewhat lackluster) attempt at an epic campaign.

So if someone is debating "Run a pre-built campaign" vs "Use AI to help me build my own", I'd argue that the former is learning less (except for the first time DMs who are learning to run the game).

What you're arguing is a bit like arguing an actor delivering the absolute best performance the world has ever seen of the opening monologue of Richard III couldn't possibly be as creative an individual as some gormless and talentless individual who just asked ChatGPT to write them a monologue.

DMs learn best through practice. The best DMs I have ever encountered have been doing it for years. Not sitting in front of ChatGPT and asking it to do for them what their own brains should be perfectly capable of doing and then incurring that 'cognitive debt' I mentioned because they lack the patience and ingenuity it takes to really learn how to do something.

I don't run published campaigns, because I don't like the pace and scale of the things, and instead run adventures that are either of my own creation or by others but that fit perfectly into what is a world of my own creation. But I would still argue running a published campaign well requires an infinite deal more creativity than we see from those who need a machine to help them come up with their own.

The latter are engaging in a practice that research has now shown harms learning.

EDIT:

What about award-winning campaigns? Things like The Enemy Within campaign for WFRP? Or Masks of Nyarlathotep for CoC? A GM who even just reads through either of these is learning how to write good games better than someone who is just feeding ChatGPT prompts. (On this point, I have evolved as a DM more from reading good fiction than anything else.)

The thing I find most amusing about AI's most ardent of proponents is they are constantly talking about how 'soon' it will be able to do this and do that. So ensorcelled are they by the 'prophecies' of the Sam Altmans of our world.

I have seen firsthand what happens when people think that way—having to listen to someone insist about 16 months ago that problems encountered in content made using generative AI would 'soon' be a thing of the past only for these problems to now be exponentially worse than they were. Everyone where I work can see it except those with an unbridled enthusiasm for AI. Who are in denial. It's like watching members of a cult insist that everything is fine when it is no such thing.

Quote from Perringaiden >>

I will add onto the "not learning" thing.

A large percentage of groups these days, run only the provided campaigns.  They don't write their own stories, and they don't want to.  They play the published campaigns, then complain when WotC doesn't provide them with new campaigns.  These are not DMs that are 'learning' beyond their first campaign.  They're reducing their workload by using a pre-built campaign, and just playing it as it's written.

Those people using AI to write their own campaigns are actually more creative than people just following a provided campaign.

It's the big difference I've seen since Basic/2nd Ed/3rd Ed, and the modern 5E game. (I didn't play 4E.)  The over-reliance on someone else to provide the story.  Back in the TSR/Early WOTC day, you'd use provide adventures in your game, throwing them in where it was appropriate and adapting them to fit.  There were campaign settings, not campaigns.  The publishers provided the world to reduce worldbuilding, so you could focus on writing the games.  These days, the worldbuilding from WotC is getting thinner and thinner, like removing the lore from the Monster Manual, because they want to focus on providing people with the next Curse of Strahd best-seller.

When you bought the Menzoberran box set, it had two books of world, and an adventure to get you started.  About the only "full campaign" I can remember was the Rod of Seven Parts, and that was their (somewhat lackluster) attempt at an epic campaign.

So if someone is debating "Run a pre-built campaign" vs "Use AI to help me build my own", I'd argue that the former is learning less (except for the first time DMs who are learning to run the game).

My experience is the exact opposite. I played nothing but pre-written modules back in the AD&D days. In 5e, some DMs will drop in pre-written adventures as side quests in their homebrew campaign (I've done it myself), but even when running something like Curse of Strahd or the new Phandelver, they were putting their own spin on it and adding in plenty of extra content

Quote from Perringaiden >>

I will add onto the "not learning" thing.

A large percentage of groups these days, run only the provided campaigns.  They don't write their own stories, and they don't want to.  They play the published campaigns, then complain when WotC doesn't provide them with new campaigns.  These are not DMs that are 'learning' beyond their first campaign.  They're reducing their workload by using a pre-built campaign, and just playing it as it's written.

Those people using AI to write their own campaigns are actually more creative than people just following a provided campaign.

It's the big difference I've seen since Basic/2nd Ed/3rd Ed, and the modern 5E game. (I didn't play 4E.)  The over-reliance on someone else to provide the story.  Back in the TSR/Early WOTC day, you'd use provide adventures in your game, throwing them in where it was appropriate and adapting them to fit.  There were campaign settings, not campaigns.  The publishers provided the world to reduce worldbuilding, so you could focus on writing the games.  These days, the worldbuilding from WotC is getting thinner and thinner, like removing the lore from the Monster Manual, because they want to focus on providing people with the next Curse of Strahd best-seller.

When you bought the Menzoberran box set, it had two books of world, and an adventure to get you started.  About the only "full campaign" I can remember was the Rod of Seven Parts, and that was their (somewhat lackluster) attempt at an epic campaign.

So if someone is debating "Run a pre-built campaign" vs "Use AI to help me build my own", I'd argue that the former is learning less (except for the first time DMs who are learning to run the game).

 
I can't say i agree with that take. When you run a prebuilt module or something Gen-AI, you are just playing what is provided.
When you had a setting sourcebook, you can reuse the world then do your own thing. When you had and adventure, you could repurpose parts. You played and grew.

Here is how it usually went.
Run published adventure, get experience, start tweaking published adventure, get experience, Buy published adventure and take a hacksaw to it, cobble it together, get experience, next published adventure, scalpel this time, craft and kitbash adventure. Next Campaign, written from scratch. 
that is why i condemned the "just run published adventures without thinking" the thinking and learning is the big part of it.
With A.I. that smashes something together, it might FEEL more creative in the moment, but.... you didn't use creative cognition to make it. You outsourced it to a server farm. You had something cobbled together for you, you didn't cobble it together yourself. Modules are like learning to ride a bike with training wheels, before taking them off to ride fully, AI is like riding cab. You didn't learn or do anything yourself.   
When you have a published adventure you can look over, or modify in the moment, you use creativity, modify to experiment. Most creative people can't help but rip it to pieces to learn.
Published modules are only ever a problem when you don't deviate from them. In my experience that isn't a lot of people. 

let me tell you something about 4E adventures. They were simple to run, simple to learn, but also.... simple to pull apart and experiment with. People were less intimidated to start ripping them to pieces for parts or to see how it worked. The encounters were very balanced and could be lifted and repurposed and in that repurposing you learn.
4E had problems, but their adventures were easy to learn from.

Quote from 7thlevelspecialist >>

 

The thing I find most amusing about AI's most ardent of proponents is they are constantly talking about how 'soon' it will be able to do this and do that. So ensorcelled are they by the 'prophecies' of the Sam Altmans of our world.

I have seen firsthand what happens when people think that way—having to listen to someone insist about 16 months ago that problems encountered in content made using generative AI would 'soon' be a thing of the past only for these problems to now be exponentially worse than they were. Everyone where I work can see it except those with an unbridled enthusiasm for AI. Who are in denial. It's like watching members of a cult insist that everything is fine when it is no such thing.

 I think you missed my actual stance on using AI then.

It can be useful, but it's not a 90% fix for your own work.  It's probably not even a 40% fix.  It can spit out a bunch of 'ideas' many of which are nonsense.  

You need to know your table.  You need to know your world.  AI can steal info on the existing worlds from things like the Forgotten Realms wiki, but so can you.  All AI is really useful for at this point is as an English major.  They can write prose, but they have no idea what it means or if it makes sense.  That's completely on you as a user.

I never claimed that you can be a "gormless and talentless indvidual", but I think you're doing a MASSIVE disservice and insult to many DMs by characterizing it that way...

Your stance is that its use is more 'beneficial' than running published campaigns when it comes to creativity and to learning how to be a good DM.

Even though—as I have said—research now shows:

(a) we aren't using certain parts of our brains when we use it to produce things for us that we do use when we create things without its help and the effect this is having on users' brains is accumulative;

(b) using it impedes actual learning. I see this every day as someone who works in education.

When neurological imaging has shown us what someone's brain looks like on ChatGPT and it does not look good, but its proponents are still falling over themselves to defend its use ... well, it is as I said before: The same research also found use of it erodes users' capacity for critical thinking.

Have you ever read, if not run, published campaigns of the caliber of The Enemy Within or Masks of Nyarlathotep? The actual mental effort that requires compared to just getting ChatGPT to give you a bunch of 'ideas' to use in your campaign is comparable to reading the texts set by one's professor compared to just getting ChatGPT to help you write your thesis because you couldn't be bothered. It's infinitely less creative to get ChatGPT to churn out dreck than it is to read good writing and process it and appropriate it for something original. Less educational, as well.

We now know those who use ChatGPT to produce things for them or even just to give them answers to things are accumulating that 'cognitive debt.'

That is now undeniable. And this is also why—like I said—it's like a cult. It's like people don't care if they literally become stupider by having machines do thinking for them they should be perfectly capable of doing themselves. But they already lack the patience and ingenuity to do that much.

Read Dune by any chance?

As for your comments regarding the amount of energy AI consumes, ecologists are not talking about how we face 'a lack of power'; what they are talking about is how we need to see a shift in attitude when it comes to consumption. The only people talking about how we just need to generate 'more power' are governments and industries ruled by artifice. The manufacture alone and then the installation of a number of 'greener' sources of energy when industrialized to meet the needs of those who just can't give up needless playthings are themselves of gross detriment to the environment.

Quote from Erik_Soong >>

Don’t do it.You will be a worse DM for it. 

 In addition to the above article, here is the article 7thlevelspecialist was citing.

AI makes your brain lazy.