I or AI?

I’ve recently been asked to talk about literature in the time of artificial intelligence (AI) in a couple of conferences in Dumaguete and Manila. What that tells me is that, with AI’s emergence and growing popularity, there’s been much uncertainty, anxiety and fear – even outright hostility – generated by the seemingly unstoppable intrusion of artificial intelligence not just into literature but into almost every aspect of human life and society. As I’ve said before, depending on how you see and use it, AI is either God’s gift to humanity or the destroyer of civilizations.

While it has been hailed for its contributions to such fields as medicine and criminology – shortening diagnostic procedures and sharpening digital forensics – AI’s application to less mechanical endeavors is more fraught with both ethical and technical questions. Studio Ghibli’s Hayao Miyazaki, for one, has forsworn the use of AI in his work, calling it “an insult to life itself.”

For writers and other creatives, the big questions are: Will and can AI replace the author? Is AI capable of artistic imagination? Should writers, publishers and readers feel threatened by its future development? Might there be a positive role for AI in literary creation?

Now, we can be very brave and declare that the worst piece of writing or art done by a human is still better than the best of what AI can produce. I’ve heard many authors proudly insist that “AI can never replace me!” But do you honestly think that’s true, and will the readers of the future – say, the consumers of popular fiction – care? The sobering fact is that there is so much bad art and bad writing done by real humans that it shouldn’t be too hard to artificially produce something better, for which people will gladly pay.

I know that this will strike some of us as being crassly commercial, but it would be naïve to deny that much of what we know to be culture today has been commodified – produced and sold as entertainment, whether it be a book, a movie, a concert, a computer program or the hardware with which to access them. These are all media in which AI is already playing an increasingly important role – initially, perhaps, merely as a facilitator, a simplifier of complex or difficult tasks or as an aid to the imagination, but also as a co-creator or collaborator, such as in the generation and animation of images.

Given the fact that most of us produce art to sell – and why shouldn’t we, especially when we promote the idea of “creative industries” – the entry of AI into our thought processes and methods of work could be a matter of survival for many. The question is, will it improve the mediocre, or degrade the excellent? Can we excel without it, or because of it?

Early AI’s clumsy mistakes or “hallucinations” are worth a laugh, but I’m not sure how long we’ll be laughing; AI’s present ineptitude simply means it has a lot to learn – and it will, with the kind of training it’s being fed off our books, our texts, our manner of writing. It will only be a matter of time – I’d say less than a decade – before AI can mimic the best of global writing, especially as literary texts get digitized and tossed into the meatgrinder, until it can produce a decent if not impressive approximation of certain styles and approaches.

(For AI professionals, the next phases of the AI revolution will move into Artificial General Intelligence or AGI, at which point AI can match human intelligence, and ASI or Artificial Super Intelligence, when AI becomes self-aware enough to improve and replicate itself without human intervention and possibly beyond human control. These scary scenarios will not take, they say, a century to happen – some experts predict that AGI could be realized as early as 2027.)

What’s going for us is that while literary styles can be copied, the human imagination is far richer and stranger than we think. AI tends to homogenize; the human artist strives to be unique. Even so, researchers are already talking about algorithmic imagination and experimental humanities as “true collaborations with culture machines.”

Given that it’s inescapable, I propose that instead of fearing it, ignoring it or maligning it as I’m sure many of us are inclined to do, we study AI and use it for what it might be able to offer in aid of the imagination – as unsettling or unappetizing as that proposition sounds.

We’re already tapping AI every time we use Google, and no one seems to mind. I don’t mind admitting that I have used AI – not in fiction but in creative nonfiction or CNF, specifically in writing the biography, where I ask AI to summarize and organize biographical material that I would have eventually found on my own, anyway – in days rather than seconds.

I suspect that the use of AI in CNF is much less troubling for writers and theorists than its employment in, say, writing the novel or the poem, which we have been trained to think of as more personal, more “us,” than nonfiction. We will yield CNF to AI, but draw the line at fiction and poetry, where we feel we should resist the intrusion of the beast or the machine into the recesses of our imagination.

I wonder, however, how long this fortress will hold, or what the first crack in the wall will be, if it isn’t there already. I’m pretty sure that somewhere out there, a plodding novelist is already using AI to chart a tree of plot possibilities – What will happen if Maria marries Oscar? What if they decide to live in Davao instead of Baguio? And so on. I wouldn’t do this myself, because the fun of writing for me is in working out the future of my characters in my head.

And then again, I write fiction for the love of it – unlike almost all other kinds of writing that I do for a living. But if I were a novelist under contract to produce a novel a year, I’m not so sure that I wouldn’t seek AI’s help to lighten my load and get the job done. So is AI OK for money but not for love? Is that what it all comes down to?

So right now, we have many more questions than answers, and at the pace the world is changing, most answers we come up with will soon be obsolete anyway. But the basic questions will remain, the most vital of which could be, when we say “I am,” is that “I” me, or is it AI speaking?

*      *      *

Email me at [email protected] and visit my blog at www.penmanila.ph.