Quote from: RPGPundit;1119789AI DMs will basically be a merging of the ttrpg and the Online RPG experience. But it will be a very long while before an AI GM is able to do everything as well as a skilled human GM can.

If ever, it may be the only thing that can replicate human intelligence is a device as complex as the human brain.

The thing to keep in mind is that they can't even build an autonomous device that is as capable as a cockroach. The day they can build a robot that can be thrown outside and survive for a few days on its own is the day to start wondering about A.I.

Until then everything is just a more sophisticated tool for humans to use. Take for example the story generation software. The core of what it does is a result of recognizing patterns in a large of amount of human written stories. Being able to recognize patterns in writing and then coherently make new stories is an accomplishment and has useful applications. However it unable to come up with new patterns on its own.  What it can do is recognize existing patterns that people have not noticed.

Voice recognition is another. It gotten very good. While annoying for other reasons in terms of recognizing your verbal responses phone tree menus have gotten noticeably better. However once you try to push it beyond what it been programmed for it is an absolute idiot.

I have to deal with this all time at a operating system in my day job which involves automation and robotic. In the thirty years it gotten way easier to setup robotics in a way that is efficient and safe. Largely due to the A.I work everybody reads about. But for all their usefulness the software is an absolute idiot when it comes to any large picture issue. That hasn't gotten better. I still have to break everything down to specifics to makes sure the setup fits the application and it is safe. It just not as fiddly as it used to be and I focus more time on the "big picture" issues.

The same with A.I. and gaming. There are patterns in what we do as referee. Software and techniques will be or have been develop that can recognize those patterns. Then be turned into a useful tool or interesting game in its own right. But it will still consistently be unable to see the big picture and be found lacking compared to a human referee.

I have to ask, at what point are you just playing a video game?

Particularly if you include text adventures as video games.

Quote from: Lurkndog;1119808I have to ask, at what point are you just playing a video game?

Particularly if you include text adventures as video games.

It is a video game. Just as not as idiotic and more flexible than the current generation. It will still be trivial to "break" the illusion in ways that a human referee can handle easily.

One thing that wasn't mentioned is the "Dwarven Forge" factor. Dwarven Forge is nice, and spectacular to use and look at once setup. But that setup is overhead in terms of time and what you need to have on hand for the game. In addition it is finite. You never have enough to represent the diversity of the setting of your campaign.

Well with any 3D graphic driven approach whether it is a human driven VTT or an algorithm driven CRPG. Unless you are 3D artist you have the same issue as you would using Dwarven Forge. For me that a major downside to any whiz-bang "next gen" tabletop RPG software/game.

One virtue of the present generation of VTT is that their 2D image approach is far more flexible than 3D.

Quote from: estar;1119796If ever, it may be the only thing that can replicate human intelligence is a device as complex as the human brain.

The thing to keep in mind is that they can't even build an autonomous device that is as capable as a cockroach. The day they can build a robot that can be thrown outside and survive for a few days on its own is the day to start wondering about A.I.

Until then everything is just a more sophisticated tool for humans to use. Take for example the story generation software. The core of what it does is a result of recognizing patterns in a large of amount of human written stories. Being able to recognize patterns in writing and then coherently make new stories is an accomplishment and has useful applications. However it unable to come up with new patterns on its own.  What it can do is recognize existing patterns that people have not noticed.

Voice recognition is another. It gotten very good. While annoying for other reasons in terms of recognizing your verbal responses phone tree menus have gotten noticeably better. However once you try to push it beyond what it been programmed for it is an absolute idiot.

I have to deal with this all time at a operating system in my day job which involves automation and robotic. In the thirty years it gotten way easier to setup robotics in a way that is efficient and safe. Largely due to the A.I work everybody reads about. But for all their usefulness the software is an absolute idiot when it comes to any large picture issue. That hasn't gotten better. I still have to break everything down to specifics to makes sure the setup fits the application and it is safe. It just not as fiddly as it used to be and I focus more time on the "big picture" issues.

The same with A.I. and gaming. There are patterns in what we do as referee. Software and techniques will be or have been develop that can recognize those patterns. Then be turned into a useful tool or interesting game in its own right. But it will still consistently be unable to see the big picture and be found lacking compared to a human referee.

Hard to say. They're getting better and better at making "fake AI" that, while clearly not actually intelligent, is able to fake it to the point of making people increasingly uncertain as to whether they're talking to a person or a machine. That's likely to keep improving to the point that it could be very effective in a relatively short span of time.