In today’s Chapter of Red Ink, we’ll be discussing the hot topic of AI in the world of content creation. Recently, there’s been a lot of debate in the creative world regarding AI. Not only is there a massive lawsuit from Bestselling authors here in the US, there is EU legislation, praised by European publishers, set to change the relationship of AI with the Publishing community. The main issue, in short, is AI has been trained on copyrighted material without consent, attribution, and compensation to the authors, artists, musicians, and publishers who created said content. If you wanna get the variety of opinions of the matter, take a look at X/Twitter. Just recently there was a big hoopla over the release of a highly anticipated horror movie Late Night with the Devil from IFC Films and Shudder and the director’s use of 3 AI created images used in the film.

The ‘LATE NIGHT WITH THE DEVIL’ directors confirm that AI is used in the film.

“We experimented with AI for 3 still images which we edited further and ultimately appear as very brief interstitials in the film.”

(Source: https://t.co/uENvCsuUiJ) pic.twitter.com/kmV6fzpqT6

— DiscussingFilm (@DiscussingFilm) March 21, 2024

Warning, I'm about to brain vomit onto the page ... lots of thoughts that may not be perfectly organized.

Going down the AI Rabbit Hole

About a year ago, it was my turn to provide a demo lesson at the language school I’m currently employed with, Telegraph Hill Language School, here in Prague, Czech Republic. I decided that I would use AI to create the lesson. I came up with the topic and started prompting ChatGPT with what I needed. All in all, a lesson plan that normally takes 1-2 hours to create, ended up taking just under an hour. Time was saved. It was a custom lesson, brought forth from my multiple prompts, that served its purpose well. It was a decent lesson. Not great, but good. I was proud at what I had created. When I shared it with the school, the director was livid. She thought it was cheating, even after I showed her the amount of time and work I put into it after refining prompts and even self-editing the content it created. In the end, she allowed it, but expressed concern that one day AI will replace us. I explained, “I’m sure that’s what horse breeders said when the car was created.” The rational floated, but it was flawed and I didn’t know it … YET.

In that meantime, I launched Red Ink blog to help with marketing of my upcoming book, Who Put Bella in the Wych Elm. I decided that to save time, I would use AI to research interesting topics, create a marketing plan, and aid me in writing the articles, so that I could focus more on writing my book. Each week I’d write up an unpolished article, ranging from 800-1200 words, and then run it through ChatGPT to clean it up, make it pretty, and have it add a little something extra to it. I also used Adobe Stock to find images to use for the banner of each post, some which were human made and some AI made. I didn’t care at the time, so I didn’t exclude either and just chose the image I thought best represented my particular post’s topic. I saw no harm in it as all the blogging software out there today is integrating AI into their systems for this exact purpose. So, everyone is doing it, right? Well, not everyone as I was about to find out.

Fast forward to the last several months when I started seeing the topic of AI appear in online writing communities and I became more aware of the current lawsuits and proposed legislation. There was so much vitriol in the writing community, and likewise in the Art community, about the use of AI and how it’s theft.

Theft? Wait, what? Did they say theft? I re-read the X posts and yes, I read them correctly. Theft. But how is AI theft? And that’s when my education into AI started. So let's dive into that.

What is AI?

Well, it isn't Hal 9000 thankfully and nor is it the Matrix. The AI systems we think of from Science Fiction movies is not the AI we have today. Those systems were near sentient (self-aware, which humans and many animals are). AI today is not. Technically they are Language Models.

A language model is a set of complex code and algorithms that consume, process, and mimic human language. The more you feed it, the better it can mimic. It observes the data sets it's fed and looks for patterns, and then the algorithm rearranges words and pixels to make something that fits the pattern. The rules of the code and the patterns in the algorithm rearrange what already exists to make something new. Therefore, the organizations that train the AI, such as OpenAI (ChatGPT) and Google (Gemini), have dumped millions and millions of pages of text (and art) into their systems for their LM (Language Model) to consume. Sure, much of it is open domain content, but they need more. Which meant they started using Copyrighted material—without permission.

But why? Why not stick to Open Domain content? There’s sooooo much of it! Because its old and antiquated. Spelling, grammar, idioms and expressions, etc would all be over a hundred years old. The LM would not sound, write, draw, or “create” like a modern person. They need modern content to train with to get modern results. And again, that meant diving into modern material which is almost exclusively Copyrighted material.

Their excuse is that Copyrighted material is “free” (to put it simply) for Educational uses, which is true for schools and universities. They see “teaching” their LM as educational, especially since they market their LM for people to do research and whatnot. But I digress.

With billions of data-sets, from both Free Domain and Copyrighted material, their LM can mimic and make content for any genre and style, which includes specific styles of individual writers and artists. Ooooooooo, there’s the another rub.

But Blake “Why did you say make and not create? Doesn’t the AI or LM create the new material?” Wow, what a good question to ask, Reader, and I’m glad you brought that up. That requires digging into what the words make vs create actually mean.

Creativity and Imagination

I can make a sandwich. I can buy all the ingredients and follow the “recipe” and get a delicious sandwich as the result. But did I create it? No … I made it. People have been making sandwiches for a very long time, there’s nothing special about that, even if I come up with different items to put on the sandwich. New ingredients are an iteration of something that already exists; it wasn’t innovative or revolutionary. It was actually John Montagu, the 4th Earl of Sandwich in 1762 who created the first sandwich. It was his imagination, while playing a card game, that came up with the easy to eat, everything in one, meal that wouldn’t interrupt his activity.

Creativity is fed by our imagination, our ability to think of things that do not actually exist. Where does that come from? As early as birth (and even the womb) our brains start to process the world around us. By the time we are 5 or 6, we are creating new worlds to play in, despite the fact we haven’t left our own bedroom. These worlds erupt from our minds because of our experiences and observations around us; the people in our lives; the stories that are read to us at night, and the films / TV we consume.

Hold up, Blake … so we are just like the LM? We just consume the world around us and our brain finds patterns, just like LM? And to that I’m going to say an emphatic, NO. And here’s why.

If simply digesting words on a page was enough, then we’d never want (or need) to travel or explore. I’ve read lots about Athens … therefore I never need to go there. I’ve watched a video of a rollercoaster … therefore I never need to ride one. I’ve been told what chocolate ice cream tastes like … therefore I never to eat any. That’s false, and humanity proves this daily. We want to see Athens with our own eyes, walk the streets Socrates and Plato walked, smell the fragrances with our own nose, feel the cold stone with our own finger tips, experience it for ourselves. Because words are not enough to replace experience. We are so desperate in our humanity to be in the world we read about that we will imagine it as kids—-and then travel there as adults: see the sights, gorge ourselves on the food, smell the flowers as they tickle our nose, and dance in the heat of a Greek summer night until we pass out or the sun comes up.

When we read a story and the author gives a description of the character, our brain fills in the missing bits to give us a full person in our mind. And no two readers will ever imagine the same character exactly the same from the pages of book, because our imagination is limitless. It has to be as it competes with the world we live in.

LM can be told these things, but it will never feel the excitement of riding your bike down the “big” hill in your town as a kid, the embarrassment of farting while giving a speech to the board of directors, the joy of a first kiss, the rejection of your first breakup, the loss of a child or parent. There are an infinite number of experiences we actually live out in our daily lives that AI never will and therefore it can never imagine even if given the words.

I love this illustration (don’t remember where I heard it):
Try explaining to an Alien who has never visited Earth, what chocolate tastes like. You can explain it a thousand different ways, using all the most illustrative descriptions, but until they actually take their first bite, they will never actually know what it tastes like. It's impossible—only experience can teach this to us.

It's this experience that feeds our imagination. Every second of your life is a snapshot moment in your brain that makes your experiences and imagination unique and separate from every single human in all of history. There has never been and never will be a person in humanity that has seen what you have seen, has tasted what you have tasted, has heard what you heard, has felt what you felt, has smelled what you smelled, and has lived when you lived, has hurt like you hurt, has loved like you loved. And no matter what you tell AI about yourself, it would be impossible for it to process every neuron in your brain and create an exact copy. It will only be a pitiful approximation.

So, LM … sadly, has no authentic imagination. Now, lets talk creativity. Our imagination fuels our creativity. The greater the imagination, the greater the creativity. But what is creativity? It's to bring forth something new into the world from your mind that didn’t exist prior. So, lets go back to the basis of an LM. It’s all about the patterns the algorithm sees and remixes the data into something—its a blender. As a computer, it follows code, the rules by which if functions. It is a slave to the rules and patterns. And there we have the big reveal.

Humans are free of rules and patterns. Sure, we have them; we even love them. But lets look at art. All the great artists in history were taught the patterns: color, lighting and shading, shapes and form, texture, styles, etc. And then you get a Picasso who says “f*ck that”; a Salvador Dali who says “what if this”; a Jackson Pollock who says “hold my beer boys and girls, check this out”. They learned the patterns of all the art before them and then said “screw the pattern” and created something new, something the world had never seen before. True creativity knows when to follow the rules and when to break them. LM can’t do that, because one) it isn’t sentient, two) it has no life experiences to draw from, and three) it can only do as the code and algorithm instructs it.

Artists of all mediums fail, and that failure, and the embarrassment that comes with it, fuels their future work. They either give up or try again, learning from those mistakes. LM is just rearranging, remixing, and making pathetic approximations of something new. If LM fails and gives us 6 fingers, it doesn't care that it failed. There is no growth or development in its failure. (click on the tweet to see the full photo to appreciate it)

How to identify AI photos pic.twitter.com/ylGkwy42fw

— TheBatmanWhoLols 🏴‍☠️🇺🇲 (@who_lols) March 27, 2024

But what if …

What if we feed so much information in the LM that those approximations become exact? Okay … but in order to do that, you will have to take what is not yours and process it. Even at its current level of copyright violation, with billions of data-sets, its still an approximation. It needs more.

As a human, I will never be able to read all the books in one genre, much less all the extant books in the world that LM can consume, yet, even in my limitations, I can create and it cannot. What kind of content could LM make if it were limited to the books I read and movies I watched? Especially considering its void of my life experiences outside of written and video media. According to AI creators, it would pale in comparison because it would not have sufficient data-sets to create accurate patarns. That and we have true cognitive abilities of rational, imagination, and creativity that LM doesn’t have. So, even if I never read a single book or seen a movie in my entire life, my experiences as who I am would put LM to shame.

Let’s do a little thought experiment. If we fed AI only music from 1899 and earlier, what would we get? Let’s assume we fed it the complete human library of music up to that moment. What would it out put? What songs / music would it make? It would only re-arrange what it currently knows. It could not and would not be able to give us music from Muddy Waters, Jimi Hendrix, Aretha Franklin, the Beatles, Lead Zeppelin, Depeche Mode, Metallica, Nirvana, Emenem, Dr. Dre, or even Nickelback for that matter. AI would only give us music similar to Antonín Dvořák or Hans Pfitzner.

Two Facebook bots where set against each other to negotiate the sale of several items. The end result? Gibberish. To them, it may have been logical in some weird way, but to us, it was nonesense. Their evolution strayed so far from what we could understand, that it became useless and nonense. Now, imagine the music.

I’m sorry, but I want my Bohemian Rhapsody on a weekly basis, not 1990s modem sounds (no offense to Dubstep fans out there).

Final Thoughts

As an author (and human), I buy books, I pay subscriptions to Kindle Unlimited and Netflix, I pay for movie tickets, all to experience the content of others—content that helps me grow. I pay for that access to feed my imagination, should not AI/LM? Should not authors be compensated? Even those like me, who is yet to publish my first book? I say yes.

What right does AI have to copy the voice and style of other authors? Should anyone off the street be able to ask AI to write a novel in the style of Stephen King or Alma Katsu and then sell that book as their own? Or even worse, those authors find passages of their works completely plagiarized in AI output without so much as a footnote (yes, it does happen).

And what if my content in the AI system is used to make something that I vehemently oppose? I don’t want AI trained from my content to be used to write Nazi or Socialist propaganda. No, and no all day every day from now to eternity. Artists today get to determine what political parties and politicians can use their music at rallies or what companies can use their music in commercials. Should I not have that same control over my content in AI?

Should AI be allowed to charge for their services and profit off their systems that were trained from my content and yet not pay royalties or give credit where credit is due? If someone stole your car and then sold it, would you be angry? Isn't that the same if they steal my copyrighted works and then pass it off as their own and profit from it?

I know AI is here to stay and it isn’t going anywhere, but the tech world doesn’t have the right to steal our work that we bled and sweat over to line their own pockets with cash. We artists of all types deserve control, attribution, and compensation when it comes to AI and our content.

Now, lets talk about me—admission time. I used AI in the past, mostly in research and spicing up my blog with style and images. While I may still use AI for random research and as a quality thesaurus, from this moment forward I commit to keeping my blog free of any AI writing assistance and from AI images (if known), even if it means I post less frequently.

Note: My book is, and was from conception, 100% FREE of AI content, including the cover art.

Post Comments

No comments for this post