This is so fascinating! I think it does have huge implications for the future development of AI in general, but I'm particularly interested in the specific case of storytelling. As a writer, I'm so anxious to find out what AI is really capable of, writing-wise. I think there's no question that AI will be capable of generating okay, reasonably satisfying stories (with all sorts of potential for things like gigantic, immersive worlds and personalized storytelling which are beyond the capabilities of a human writer), but will AI ever be able to create GREAT stories? Or is there, finally, a human element that is indispensable and irreplaceable -- something along the lines of true originality or genuine emotional depth? And will AI help a great writer become even greater, or will it just dilute his genius? We may want to believe this or that, but I don't think anyone really knows the answers to these questions. Not yet.
Great question! Of course, I have no idea, but I'd like to believe that what makes great stories is having great lives (I think Hemingway said so?) so without embodiment, without being there in the world, I don't think AI can write a truly great story. Perhaps when we have machines that are embodied in physical form and can go to wars and fall in love and be betrayed and actually suffer... I think that's a unique part of being human (or a living, sentient being in general). Perhaps if a stray dog or cat could talk, it could tell great stories.
I’ve noted the name and function of each agent ranging from rule extraction to knowledge graph to player action handler. The metacognitive mapping here on the part of the handler seems semantic frame related linked to the knowledge graph. Most of the others seem keyed to narrativity. As it should, I mean, given your interest in stories.
I’m trying to look at different level of abstraction for agent relationships that could handle expository structures, particularly theory building from empirical data. Methods section, Findings, Interpretation, Discussion. I’ve long been told a good study is a good story.
How do we (people like me) get our hands on AI systems like this? I literally have no idea.
Hey Terry. One thing is we definitely need more humanities researchers in these projects, to help conceptualizing what good stories look like, or even if those are meaningful concepts and questions, and of course in terms of evaluating not only the end result but the intermediate steps.
Otoh regarding your last question, so far this is pretty much research-level code which means there's no user friendly app or anything resembling something you could use, sadly.
But we can work together on this, perhaps establish a joint research project (even an informal one) and my students could share internal tools you and other interested people could try and use. That kind of collaboration would be something very very interesting to me.
I’d very much like to get involved in some useful way in this project. I would have a steep learning curve but I’m a good student when I’m motivated. I think Nick Potkalitsky would love to contribute to this line of research as well. Nick’s doctorate is in rhetorical narrativity. I need to talk to him. Right now we are finishing the final draft of a book and will be snowed in for a month or so but then your invitation could be the most important thing we could do.
Perfect! I'd love to have Nick on board. I'm wrapping up the course as well but next semester I'll be back with novel research on interactive storytelling so we can use summer break to make some plans and begin next semester with a more or less clear path forward.
This is so fascinating! I think it does have huge implications for the future development of AI in general, but I'm particularly interested in the specific case of storytelling. As a writer, I'm so anxious to find out what AI is really capable of, writing-wise. I think there's no question that AI will be capable of generating okay, reasonably satisfying stories (with all sorts of potential for things like gigantic, immersive worlds and personalized storytelling which are beyond the capabilities of a human writer), but will AI ever be able to create GREAT stories? Or is there, finally, a human element that is indispensable and irreplaceable -- something along the lines of true originality or genuine emotional depth? And will AI help a great writer become even greater, or will it just dilute his genius? We may want to believe this or that, but I don't think anyone really knows the answers to these questions. Not yet.
Great question! Of course, I have no idea, but I'd like to believe that what makes great stories is having great lives (I think Hemingway said so?) so without embodiment, without being there in the world, I don't think AI can write a truly great story. Perhaps when we have machines that are embodied in physical form and can go to wars and fall in love and be betrayed and actually suffer... I think that's a unique part of being human (or a living, sentient being in general). Perhaps if a stray dog or cat could talk, it could tell great stories.
oh boy! ai slop!
i can’t hardly wait
to retch.
I’ve noted the name and function of each agent ranging from rule extraction to knowledge graph to player action handler. The metacognitive mapping here on the part of the handler seems semantic frame related linked to the knowledge graph. Most of the others seem keyed to narrativity. As it should, I mean, given your interest in stories.
I’m trying to look at different level of abstraction for agent relationships that could handle expository structures, particularly theory building from empirical data. Methods section, Findings, Interpretation, Discussion. I’ve long been told a good study is a good story.
How do we (people like me) get our hands on AI systems like this? I literally have no idea.
Hey Terry. One thing is we definitely need more humanities researchers in these projects, to help conceptualizing what good stories look like, or even if those are meaningful concepts and questions, and of course in terms of evaluating not only the end result but the intermediate steps.
Otoh regarding your last question, so far this is pretty much research-level code which means there's no user friendly app or anything resembling something you could use, sadly.
But we can work together on this, perhaps establish a joint research project (even an informal one) and my students could share internal tools you and other interested people could try and use. That kind of collaboration would be something very very interesting to me.
I’d very much like to get involved in some useful way in this project. I would have a steep learning curve but I’m a good student when I’m motivated. I think Nick Potkalitsky would love to contribute to this line of research as well. Nick’s doctorate is in rhetorical narrativity. I need to talk to him. Right now we are finishing the final draft of a book and will be snowed in for a month or so but then your invitation could be the most important thing we could do.
Perfect! I'd love to have Nick on board. I'm wrapping up the course as well but next semester I'll be back with novel research on interactive storytelling so we can use summer break to make some plans and begin next semester with a more or less clear path forward.
Sounds great. I’ll be in touch. Thanks