What If… AIs write better stories than us?

Wall-E writing

I’ve recently started a new series of speculative posts, in which I pose a What If question and try to investigate its potential. This is an open discussion and you’re all very welcome to participate. Or just teleport the discussion to your own blog. Or to the far reaches of outer space where it can radiate in all its glory, and make all nearby stars go supernova. Explosions are so much FUN! Blaze, you nifty little idea gremlins, blaze in galactic fire! Mwahaha!!

*I really shouldn’t drink Red Bull in the morning*

 

What If… AIs are better storytellers than us?

Just thinking of the machines in The Matrix and their ability to create a perfect simulation of reality, makes me green with envy. Isn’t that the ultimate goal of all writers? To create fictional worlds that feel so real, so compelling that our audience gets completely lost inside of them? What if the artificial intelligences we will inevitably create one day (or fail to prevent from emerging) will be so masterful in creating false realities that they’ll completely outstrip us?

And what if our crazed paranoia about the death of the publishing industry is, in fact, the ultimate doomsday prophecy? *dun-dun-duunn*

Robots that write books have already been designed, and are fairly successful too. From Plotto the master book of all plots that was turned into a veritable plot-churning machine, and software that writes online articles, to Lambert M. Surhone who (or should I say that?) apparently produced over 150,000 books, machines have proved their potential for mastery of the written word time and again. One even wrote a poem that I must admit [expand title=”left me wondering…”]

To Truth

To truth I offer this thanks,
when needing something like reality
When I’m writing and drawing blanks,
I almost settle using actuality

I am in search of more,
trying to sing your praise!
It’s you I very much adore,
lacking in so many ways.

(source: BBC)

[/expand]

But robots are still very far away from being truly creative the way we scribblers understand creativity. They don’t incorporate life-experiences into fictional characters in new and enlightening ways, and don’t offer us a wider, deeper understanding of ourselves.

“I couldn’t think of anything more pointless than reading a piece of fiction written by a robot. […] You might find a lot of regurgitated platitudes but I can’t imagine a piece of software being capable of producing something that would stop you in your tracks. Not until we get truly intelligent computers.” ~Alastair Reynolds

Yeah… so what about AIs?

What if AIs with their human-like or uber-human intelligence will write better stories than us? Speculative fiction that makes our imagination look positively obtuse? Literary novels that exceed our understanding of our own humanity? Poetry, discovery novels, movie scripts and animations, adaptive games and trans-media storyworlds that take the very concept of storytelling to a whole new level?

It makes sense that artificial intelligences will be able to grasp the concept of fiction and distinguish it from fact (after all, one can be cross-checked with several other types of information sources, one cannot). Creating fiction is the next logical step. And why wouldn’t they try to do it? If they have the computing capacities to run an autonomous intelligence, they also have the capacity to run a representation of the world — and fiction takes the same amount of resources as reality. It’s all just data, but fiction has the added bonus of keeping humans happily engaged in a world they can’t control, while the AIs slowly take over the real world. Or something like that.

But let’s say they’re not out to annihilate us (that speculative concept has somewhat gotten old over the years anyway). What would their stories be like? Will their science-fiction be as conflicted about the path we’re taking as a species as ours is, swaying from enthusiasm to apprehensions and back again? Will their fantasy be just as fascinated with myths and folklore as ours? Would their horror be equally preoccupied with unnatural mutations of either mind or body? Would their romances be about beating impossible odds for the one true love? Would their crime be as ruthless in exposing the dark side of society?

Or would they not write about human concerns at all?

They might write about their own fears and hopes instead. Maybe their novels will be about an existence without physical limitation, or about having an eternal consciousness that will only know continuous growth but no end? What might love be to them, and hope? And what in the world would they be most afraid of, being shut down or becoming limited? Will they have a muse, a creative process? Will they need drafts, or would they write perfectly from the start (or not even think there’s a need for perfection)?

Maybe they will write about the hardship of putting up with our limited minds and our idiosyncrasies. Will they speculate about how our minds and emotions work, just like we speculate about theirs? “Do Bipeds Dream Of Organic Sheep?”

I think there’s something intrinsically self-reflective and philosophical in the need to create art and fictive realities. It’s one of the most effective ways to understand ourselves, and give our lives meaning. If AIs are as fascinated with fiction and creating their version of it as we are… I believe we couldn’t ask for a clearer sign of a higher intelligence.

What do you think? Will AIs play with words? And will they do it for us, about us, or despite of us?

13 Replies to “What If… AIs write better stories than us?”

  1. A great article, Vero and it struck a chord here:
    Creating fiction is the next logical step. And why wouldn’t they try to do it? If they have the computing capacities to run an autonomous intelligence, they also have the capacity to run a representation of the world — and fiction takes the same amount of resources as reality. It’s all just data, but fiction has the added bonus of keeping humans happily engaged in a world they can’t control, while the AIs slowly take over the real world. Or something like that.
    Because this is a key aspect of my latest WIP. Ref what they would right about/create to keep humanity happy? Porn and romance. A bleak future perhaps but to paraphrase someone wiser than me. ‘We get the rulers we want’

    Like

    1. Porn and romance for sure. I don’t think anything is as entrancing to the human mind as the satisfaction of its most primal needs, bonding to another and reproduction.

      I always wondered why the Matrix was as “normal”in that sense as it was presented in the movies. I don’t buy into the explanation that it needed to be as realistic as possible so people won’t notice they’re trapped. Most people don’t want to be living in reality, and for those who do, what’s to stop the Matrix machines to give them a different, tailored reality to live in?

      Like

  2. Interesting.

    Whenever I see “AI” in the name, I imagine something truly opposite of what futurists and transhumanists envision it to be. Having spent my better days in computing field I would strongly advise philosophers and abstractionists to stay away from machines. You all are wonderful people, but NO.

    The thing is, the AIs probably will be anything but what we imagine them to be. You wouldn’t want to write a set of programs that will spiral into existential loop and then, maybe, existential crisis. And I wouldn’t want my AI suddenly debating my vital decision in poetic form on the basis of some strange philosophy it might have developed overnight while I took my nap. Now imagine it runs a ship, far away from home, on a mission. I’ll be damned to have that onboard. I’d kill it. 😀

    All this brings us to the topic of what concepts we would want to share with the AIs. We would want them to be efficient as well as creative, so fiction is fine, but we better teach them to distinguish between the reality and fiction first. Seriously. 😀

    That would bring us to this dreaded concept of how AIs would percept reality and what will reality be to them.

    Well, smth like this.

    Like

    1. If you haven’t read “Destination:Void” yet, you must. 🙂

      Thing is, what if you can’t teach an AI anything in the classic sense? What if after it becomes self-aware, you’re written off as an unreliable source of information? Did you trust anything your parents tried to teach you when you went through puberty? I certainly didn’t. And I expect an AI gaining self-awareness to “feel” pretty much like a teenager discovering he’s got a personality distinct from all else around him, and that he’s got a right to say NO.

      Like

      1. Yes, I’ve read Herbert’s D:V. No, I wasn’t thinking of this, but close enough. LOL

        When I think of “teaching” AI, I don’t imagine it to be in the classical sense at all. It will possibly be more of a self-study, but you wouldn’t want to give it full open access to all of your data, would you? 😀 (All your bases are belong to us. lol)

        My point is, I wouldn’t build an AI like a teenager. That’s why I said this ‘being’ would be totally different from what we imagine it to be. We behave as our brains had evolved and developed to behave. An AI is different in that sense, and I wouldn’t base its processor/brain on a human-type brain and the overall chemistry surrounding our decision making and shit, because of how we process information. That’s not a good model. Maybe self-awareness is not what we should aim for at all, and it is possible that it might never occur to an AI.

        Like

      2. I’m fairly certain there is a threshold that when crossed, any sufficiently complex system will realize it is distinct from its environment and its observations, and thus become self-aware. However, if we can call that a “personality” is highly questionable.

        But AIs with a different way to “think” than the way human brains think (process information from the senses and the own body) pose an interesting notion. 🙂

        Like

  3. Years ago there was an AI psychologist program. I think it was called Lisa. A person would type a line and Lisa would give a response. It might go something like – I feel sad. Lisa would then say, Why do you feel sad? It was programmed to discourage swear words. And sometimes the responses clearly indicated that it was an AI program, not understanding the depth of human emotions. I suspect AI authors would write novels that would not portray the depth of human feelings and emotions. Of course there could be a new genre 🙂 AI’s writing for AI’s.

    Like

    1. AIs would probably redefine genre in a very algorithmic and deconstructive way. Fiction-science, mechanicism, recalibrational ballads, maybe even turn poems into system-altering pieces of code that carry dangerous new protocols between their rhymes. :S

      I can’t imagine an AI as shrink or counselor… what advice could he give to a being he can’t empathize with?

      But then again, humans would make poor shrinks to AIs as well…

      Thanks for the thought provoking comment, Heather! Now I’m off on several tangents at the same time! 😀

      Like

  4. I’m afraid I must apologize, Vero.

    I had typed up one of my typically long, rambling comments, then realized about halfway through that what I was typing was a great idea for a story. So I erased it all and copied it into my “idea nuggets” folder, not to see the light of day until I ripped it apart and spilled its guts all over an empty page. Your loss is my gain, I’m afraid!

    Still, great post. And thanks for the inspiration (again)! 😉

    Like

  5. My first (abandoned) novel was about an android who started sculpting and painting. Her works were odd and mathematically obtuse, and she had to explain them to the humans she showed them to. But when other androids started seeing the work, they were deeply moved.

    When/if we get true AI’s, I don’t see them writing stories for us any more than we’ll try writing stories for them.

    Either that, or they’ll have been built as cheerful slaves, happy to entertain us as we drift along in Pixar’s Axiom.

    Like

    1. Hi, Dan! That sounds like an interesting story. Maybe you’ll pick it back up one day. 🙂

      If AIs create art, and have a thorough understanding of what purpose it serves to them, I’m fairly sure they will create art for each other, rather than for us. But I also think we’re going to try to adopt them into our schools of thought and philosophy, and maybe even create some amazing art together, especially plastic art. Not so sure about storytelling, though… there might be huge differences in our concepts.

      But isn’t that the purpose of stories, to build bridges between different inner worlds? 🙂

      Like

Comments are closed.