This week: AI: Some Personal Opinions Edited by: Max Griffin 🏳️🌈  
More Newsletters By This Editor 
 
1. About this Newsletter
2. A Word from our Sponsor
3. Letter from the Editor
4. Editor's Picks
5. A Word from Writing.Com
6. Ask & Answer
7. Removal instructions
| My opinions about authors using AI have evolved over time. This newsletter gives some personal thoughts about AI and where we might be headed. |
I’ve written previously about AI and copyright, but this time I’m writing to give some personal perspectives about this subject. My opinions on AI are evolving—as opinions should. Ideally, opinions have multiple inputs, ranging from emotional to experiential to numeric. In the case of AI, the history of pervious technological innovations and their impact on artists is also relevant.
The point is that what you’re getting here are my opinions. This is quite different from my earlier newsletters on this topic, which were largely reporting findings —present and historical—of US case law and of the US Copyright Office. Those findings, especially those of the Copyright Office, have weighed heavily on my opinions, but so have other inputs.
What Are the Questions?
I’ve got lots of opinions about lots of things, but this newsletter is about a set of specific questions related to AI and authors. Here’s my list.
 Will AI-authored fiction replace human-authored fiction?
I’ll answer that one right now: it’s an emphatic no. There’s more detail later, but, well, no.
Will AI-authored fiction replace some human-authored fiction?
I can answer that one right now, too: yes. I just can’t predict what kinds of fiction will be at risk. Again, more detail later.
Is AI-generated fiction art?
How might an author legitimately use geneartive AI?
Who benefits from the use of generative AI?
That’s enough for one newsletter. But before going on, I want to say what this newsletter is not about. It’s not about the multitude of ways that AI could—and already is—being abused and misused. It’s not about the threats that AI presents to society as a whole, or to segments of society, or or the economy, although I do make some comments on these things. It’s not about how AI needs to be regulated. Indeed, my level of trust in our current governmental regulators is at an all-time low right now. This newsletter is only about the above five questions.
Will AI-authored fiction replace human-authored fiction?
I’ve already said the answer is no. Here, the example of photography is helpful.
In the nineteenth century, prior to the invention of the camera, many artists made a living doing portraits. Today, there are still a few artists who make a living painting portraits—I have a such a painting of my daughter hanging in my office, done by a local artist in my community. But I also have many photographic portraits of her in my office. So, what’s clear, is that photography hasn’t completely replaced people who paint portraits, but it has reduced the number of such artists. It has replaced them with a new kind of artist, one who takes photographic portraits.
Now, it’s true, photographic portraits are different from painted portraits in many ways. The typical photographic portrait doesn’t match the work of, say, Rembrandt, but the same could be said of the vast majority of historical portrait artists. Moreover, almost no one in the nineteenth century could afford to hire a portrait artist, while today portrait photography is widely available. I have no idea what my namesake and great-grandfather looked like, but my grandsons will know what their great-grandfather looked like since I have a photographic portrait of him.
The example of photography makes me certain that AI-generated fiction won’t completely replace human-generated fiction.
Will AI-authored fiction replace some human-authored fiction?
I’d have to say the answer here is yes, but I can’t say exactly what fiction might be replaced. Formulaic fiction is the kind that comes to mind. The history of the the Nancy Drew and Hardy Boys series is instructive here. I could readily see the assembly-line approach that the Stratemyer Syndicate used to produce these books being replicated by AI engines, for example.
Other formulaic genres come to mind. There’s one in particular, but this is a G-rated newsletter so I’ll move on to next the question.
Is AI-generated fiction art?
Well, I’d have to say my personal answer to this question is no. For me, art is a uniquely human activity. It’s an expression of our human urge to create, to imagine, to communicate, to touch another soul. Creating art involves thinking about other people, about the world, and our collective place in it. A machine can imitate art, but it can’t create art because it can't think.
As it currently stands, AI is just a computer making predictions about what comes next based on a prompt. That means it's looking at a string of words and deciding what string of words is most likely to come next and also sound like human language. The predictions aren’t creative by themselves—they are imitative. AI is NOT answering a question, even though it sounds like it is. It is NOT searching for links--a prompt can give you nonexistent links and can give links with incorrect answers. It is merely predicting the next set of words based on its training data and the words in the prompt. It's looking for similar strings of words in its training data and then producing a new string of words by "sampling" the similar strings and then constructing something that sounds like a human answering a question.
Now, it’s true that what I write is influenced by what I’ve read or otherwise experienced, and in that sense it's also imitative. But it’s also got a creative spark that’s part of me in it: I haven't copied text from anyplace, I've learned from it. There’s no me, no conscious entity, lurking inside our current versions of AI. There might be someday, but before that happens we’ll have to figure out what it means to be self-aware, to be conscious. That’s a deeper problem that philosophers through the ages have struggled with and is still not resolved.
Thus, the present state of what we call AI can simulate art. It probably can even evoke emotional experiences that resonate with humans in ways that are similar to real art. But, by itself, it’s not art, because there is no self-aware being who created it.
On this one, I’m certainly open to being wrong, but I’d need a more convincing test for self-awareness than the one Turing proposed, or even the one Kant suggested.
How might an author legitimately use AI-generated text?
There are few ways right now that occur to me. For example, people who are neurologically divergent are already using AI in creative ways to produce art. This use makes creating art more accessible to more people. This is a use I can enthusiastically support.
If one thinks about this, everyone is divergent in varying ways and degrees. Maybe authors could use AI to expand the reach of their experiences, or at least expand their personal training data.
For example, I sometimes write about fictional planets around fictional stars. Obviously, I’ve never experienced being on such a planet, but I do want to get the details right. So, for example, I do research to be sure that I’ve gotten things like the color of the sky correct. That research involves Google searches (which, alas, are now AI “assisted”—AI is NOT intelligent and can and does make errors of fact), but it takes me to legitimate academic sources. I’m expanding my own personal training data when I do these searches.
Now extend this idea a bit. My personal training data, for example, doesn’t include what it feels like to jump out of an airplane and, god willing, it never will. But the training data for AI includes people who have written about that experience. Why not use AI to sample that data for me? Is there anything wrong with using what I learn from that in my own writing?
I admit that there could be something wrong with that on several levels. AI is essentially copying from its training data. If that data includes copyrighted works, it’s certainly wrong. Indeed, since generative AI is copying stuff from its training data, copying AI-generated text word-for-word into my work satisfies the definition of plagiarism, so, yes, it’s wrong. But it’s not wrong to learn from what AI produces and use it to inform what I write. The difference is subtle, but important.
Here's a way I’ve personally benefited from AI. I find it helpful to have a photograph of my characters when I’m writing a story. I used to scroll through photograph sites looking for commercially available images that more-or-less matched my mental image of my characters. I recently took the description I’d written of my characters in a story and used it as a prompt to produce an AI-generated image. It took several iterations of my prompt, but after a few minutes I had quite good approximations of my mental image of the characters. You can see the results here:
I don’t expect to ever publish these pictures, but they will help me pick up this sequence of stories when I resume writing them—like after I finish writing this stupid newsletter.
I'd never ask AI to evaluate my writing or make suggestions on character, plot. or craft. Remember, it's NOT "thinking" about my question. It's looking for strings of words that might follow whatever is in the prompt, with no attention to the accuracy or even the meaning of the response. It might say my writing sounds like Hemingway to one prompt, and, with a slightly different prompt, it might say it sounds like an illiterate ten-year-old. Don't ask generalized AI for information or evaluation, at least if you care about the accuracy of the answer you get.
In contrast to generized AI, specialized AI, for example for medical diagnoses, can have specialized training data. In the case of diagnosing based on a list of symptoms, it can respond with reasonable accuracy with a list of possible conditions. To the best of my knowledge, there is no reliable "literary critic" database that can be used to evaluate the merits of fiction.
I’d like to return for a moment to the example of photography. Remember that photography created an entire new art form. We not only have artists like Annie Leibovitz and Richard Avedon doing portraits, we also have artists like Ansel Adams or Richard Misrach doing landscapes or Alfred Eisenstaedt’s iconic photo of the VJ-day kiss . No one could have anticipated this new art form at the beginning of the photographic era.
We’re at the beginning a similar new era today with the launch of generative and agentic AI. No one can predict where we’ll be in ten years, let alone in the next century. In my opinion, AI has enormous and exciting potential for new art, new artists, and new ways for humans to be creative.
Who benefits from the use of generative AI?
As the prior section makes clear, it’s my opinion that authors can benefit from AI-generated text. That’s not only true for neurologically or otherwise divergent authors, it’s true for all authors.
In particular, generative AI can enable more authors to reach more readers more effectively, so everyone wins in this process: authors and readers. Readers benefit not only by having more things to read, but through having more viewpoints and experiences to sample.
Note that I’m not saying everyone will benefit financially from this. Some people will certainly not benefit financially. An example is law firms that are already replacing clerks with generative AI for routine tasks, reducing the employment opportunities for not just clerks but lawyers. In the process, the law firms are making higher profits. There are winners and losers, and no one wants to be a loser.
It’s my opinion that history shows it’s futile to fight a transition to new technology. However, it’s also my opinion that making the transition equitable is a legitimate goal.
New technology is, in the long-run, almost always disruptive and introduces inequities during the transition. The problem today is the pace of the disruption is so much faster than in the past. It took decades for photography to displace portrait artists; today it takes a few short years or, sometimes, just months for the disruption to run full cycle.
Our economic systems are ill-suited to equitably distribute the costs and benefits of this kind of rapid transition. Both the costs and the benefits are likely to be high, which makes this a high-stakes endeavor. However, see my comments above about the current state of potential regulators. The disruption from AI has the potential to be so pervasive and so severe that systemic collapse is possible--see black swan events. . While I’m excited by the artistic prospects that AI presents, the overall event horizon is much less sanguine.
I’ll close on that note, along with an invitation for your comments.
|
Have an opinion on what you've read here today? Then send the Editor feedback! Find an item that you think would be perfect for showcasing here? Submit it for consideration in the newsletter!
https://www.Writing.Com/go/nl_form
To stop receiving this newsletter, click here for your newsletter subscription list. Simply uncheck the box next to any newsletter(s) you wish to cancel and then click to "Submit Changes". You can edit your subscriptions at any time.
|
For Authors: December 17, 2025 Issue [#13500]
View Portfolio
Visit Notebook
Send Gift Points
Awards
Badges
Send Email
Community Newsfeed
Puzzle Hub
General Discussion
Sponsored Items
Book of Trinkets
Static Items
Books
Groups
Interactive Stories
Audio
Campfire Creatives
Community Notes
Crossword Puzzles
Documents
Folders
Images
In & Outs
Madlibs
Photo Albums
Polls
Product Reviews
Quizzes
Survey Forms
Web Pages
Word Searches
