For some months now I have been thinking, ‘“Wouldn’t it be cool to generate AI photos and videos of my characters? Seriously! I mean online fiction like mine is difficult to read even at the best of times and could perhaps need some illustration, right?
I have also always been bugged by the fact that ebooks were just that – a digital clone of a paperbook, and from a creative standpoint if you do digital storytelling, the ebook just comes off as a poor man’s version of it. Hence why I have my website. It contains a lot more possibilities for creative expression and I am an artist before I am a storyteller (or at least that is how I like to think of my own humble self … pfff).
But anyway, this means I “ought” to do something niftier, right? Especially with – yum, yum – all the AI possibilities out there … right?
The problem is that it may not work … it may not make my stories better.
Nope. Reading about my characters may not be a better experience if you get pictures or video clips of Carrie or Jon or Marcus Chen acting out some of the scenes or maybe doing new scenes, to sort of cap off the actual story. Really.
Or maybe Carrie and Lin taking a polaroid selfie (or something) one a fine day in 1996? Just a random thought …
Anyway, let me try to explain why I think many character illustrations for the stories, AI or not, may be a shot in the foot …
1) It Ruins Immersion
My primary concern with using AI-generated images of characters in my fiction is that it fundamentally disrupts the reading experience. Reading is an immersive, internal process—you’re in your head, imagining the characters. When you insert photos of characters, whether AI-generated or real actors from a movie, you’re asking readers to do two contradictory things simultaneously: immerse themselves in the text and process visual information.
So yes, it would have been the same if I used photos with real models (if I had access to that) – people who specifically were to portray, say, Carrie and Jon, and who would be the same throughout the site. (Or old-fashioned illustrations drawn in some kind of semi-realistic style.)
It all challenges immersion!
The symbolic images I currently use as featured headers are different – or at least I consider them different. They’re not ubiquitous throughout the text, (usually) just a single image at the top. They’re clearly symbolic—reminiscent of the character’s situation without claiming to be the characters—and they function more like a book cover: setting a vibe, striking a tone, and then fading from consciousness as you dive into the story.
2) It Contributes Nothing New
Beyond the immersion issue, photorealistic depictions of characters simply don’t add value to the reading experience. I’ve always disliked illustrations in classic novels for this reason—they feel superfluous.
I mean, really. If I tell a story about Carrie and Jon having an argument, do you need to also see a photo(realistic depiction) of them arguing at the same time you are reading?!
Your job as a reader is to immerse yourself in the text. Illustrations of this kind only serve a purpose in nonfiction, where they clarify information (like a battlefield map) rather than compete with imagination.
I mean, either you do a book, or you do a movie, right? I’ll get back to this …
3) It Creates Distractions
This is a variation of the first point, I guess. But the difference is that this time it’s about a potential “threat” to immersion, not the actual ruining of immersion. Maybe it’s semantics, but I thought I’d do this in a separate point. The trick here is NOT to create more “clutter” that can potentially distract the reader. And this is what AI pics with online fiction stories might do, because …
Inserting pictures or videos into prose creates a fundamental confusion about what medium you’re engaging with. It’s like watching a 4D movie where they spray water in your face during an action scene—suddenly you’re reminded you’re in a theater, not immersed in the story world.
Most people are conditioned to either read or watch something. You’re effectively asking them to immerse themselves in two different media simultaneously, which fragments the experience.
Even if the AI generation is technically impressive, readers can usually tell it’s AI—especially given the context. I’m just a writer; I haven’t hired models or adapted my work into a film. So when readers see these images, they’re thinking, “This person isn’t real. She looks real, but she isn’t.” That’s a meta-level distraction that pulls them out of the story entirely.
Oh, yeah and some readers may have ethical concerns about AI-generated content. While that’s not the discussion I’m interested in having here, it’s another potential layer of distraction from the actual story. My focus is on whether AI serves the creative vision—and currently, it doesn’t.
So, all of these might only damage the reading experience. Or they may not. There are no guarantees and there can be valid reasons for using AI pics to illustrate your story online, or just other types of trade-offs or cost-benefits. Sure. No prob. But these are the ones that hold me back from going further with this … now.
*
Aside from all of that, there’s also the mundane issue: I barely have time to write. Spending late nights fiddling with video generation would be untenable regardless of these other concerns. Just sayin’ …
*
What About the Future?
Will I eventually use AI to tell my stories? Almost certainly—but in a completely different context.
When it becomes easy to generate, say, a 30-minute AI TV series episode equivalent to one of my short stories, I would absolutely do it. At that point, it would be my creative vision rendered visually, no different from if I’d created a comic book (I can draw, I just don’t have the time). The software would presumably be sophisticated enough that storytelling wouldn’t require endless technical fiddling.
The key difference: I would put AI video content on a completely separate platform—probably YouTube or whatever is posh at the time. When viewers go to that platform, in this case the Tube, they’re already in “video consumption mode.” There’s no text to negotiate alongside the visuals, no competing mediums. The only adjustment is accepting that the characters are AI-generated, and I suspect that in a few years, most people won’t care about that distinction.
The Core Principle: Don’t Mix Forms
The tension in my current situation isn’t about whether AI is “fake” or unethical. It’s about asking people to have the wrong kind of consumption experience.
Online, there’s a temptation to jazz up text with videos, photos, and multimedia elements. Like I said, above I think ebooks feel boring, and many websites still feel static. So I have and will continue to experiment with that, but more like ‘asides‘, elements that are supportive of the experience of reading, not disruptive (at least I hope not). And I will continue to experiment with whatever I can of the tech available to me, when making a website.
So there will be more symbolic header images, a few headlines and other textual enhancements, maybe even period-appropriate YouTube videos at the end (like a 1995 movie trailer the characters might have seen), or similar. That will not go away or change. I will try to put more of it in!
And I will do it because these elements function as contextual artifacts rather than literal depictions. And you can easily skip them when you read. They also help break up the text without distracting (unduly) from it. At least that is the idea …
But even though I am sorely tempted some days, you are not going to see AI-depictions of Carrie, Jon and co. (or even models specifically hired for the purpose, if I won a million dollars!).
I believe that the moment I start inserting character depictions into the text itself, scene depictions, video-versions of key scenes – then I force readers to negotiate the reading experience and not just immerse themselves and enjoy it as it is. And that seems … not right.
Anyway, that’s how I feel about this today. We’ll see how I feel about it tomorrow. Perhaps there will be some shiny new AI or other-thing that will change my mind … Eventually, it will probably happen. The only question is when.
*
New stories …
*
So here’s a selection of what I have been writing since last … feel free to check out anything that resonates with you (and let me know what you think!)
Strawberry Fields in waiting …
A flash piece about Deborah’s time in New York City, searching for success as a poet and artist but finding mostly low-paid jobs subbing as teacher or waiting tables. Year: 1976
Gangsta’s Paradise, Backstreet Boys and Ace of Base …
A new piece from when Carrie lived with her mom, Deborah, in a run down papermill converted to condos – in outskirts Cleveland. Mid-1990s mood piece that sets the tone for the ambiguous relationship between Carrie and her friend, Lin, who later committed suicide. Year: 1996.
On the Road to Nowhere?
A little mood piece near the end of Carrie’s dark years when she was an addict, roaming the roads of North and South America in search of redemption – and just survival. Year: 2004.
More stories in this sequence coming soon!
You might also enjoy these longer works from that era: That Which Cannot Be Broken, Scars of Our Civil Wars, and Shadows in the Shape of Men
Life Fractures
Three small pieces about Carrie, her mother and stepfather in the late 2010s. About feeling trapped as a special needs parent, getting older and that eternal search for a pivot in life (and why you may not be able to). Year: 2017
That’s it for now. Let me know in the comments if any of the stories resonated with you!
I will see you in the next Shade of the Morning Sun-newsletter!
*
Photo credit for this newsletter’s header: Google’s Nano Banana.
Share a Thought