The age of AI-generated content is here
Chances are you’ve read something written by a bot and didn’t even notice
--
A headline catches your attention. You read the article with low expectations, but the sentences flow with ease — the author is speaking your language. The ideas resonate with you. You create a connection with the author. That’s it, this person is talking to you. There’s a human connection happening here. What you don’t know is the article has been written by an AI.
Let me assure you of two things:
- The person writing the words you’re reading now is very much a human. I’m in my bed with my laptop — I needed a break from the desk — looking through the window at the spectacularly green trees of Vancouver’s spring and thinking if I should finish writing this article outside.
- A text-generator AI could’ve written this last sentence.
Yes, it’s 2022 and the future is here: we’re entering the age of AI-generated content. Don’t think AI models are there yet? Google “AI copy” and you will find businesses already making money out of it. These bots might not be at the point where they can write full-length novels yet, but here are a few things AI can successfully do (especially when well-curated):
- blog posts
- tweets
- poems
- marketing copy
- customer-service chat
- YouTube scripts
Chances are you’ve read something written by a bot and didn’t even notice. Writing bots are here and it’s just the beginning. In the age of content creation, it’s a lot easier (and cheaper) to have a bot come up with ideas and write you drafts than having to come up with drafts yourself or having a team of writers do it. It makes the task of creating content that will resonate with your readers a lot easier.
As a reader, then, the question becomes: when that connection happens, is it real? Are you connecting with the bot? Are you connecting with the person curating the bot’s creation? Can the curator even take credit for the bot’s creation?
Artificial Intelligence is not intelligent (yet)
Let me get this out of the way before you start getting worried: writing bots aren’t self-aware. All they do is replicate human ideas and styles after seeing millions or even billions of samples of them. They have been trained to solve an equation, and when you ask them “what’s 2 + 2,” they will answer 4. If you ask “how do you like your coffee?” they will answer “coffee should be like movies: dark and make you feel something inside” — although they have never had any coffee nor watched any movies.
When a bot says “What’s even the point of living if we’re all gonna be replaced by machines?” it might resonate with you, but the bot doesn’t know what it’s saying. It doesn’t know what “living” is or what the word “we” even means.
If you ask a bot what life is it will give you a satisfying answer, but it can’t internalize the concepts it’s talking about.
The bot is just repeating what a human would say. It’s like when a child says something only adults normally say — except instead of being cute and funny it’s creepy.
Are you in love with a bot?
If you read a poem about love written by a bot and love it, are you connecting with the bot? I’d say no, as it has no idea what it’s talking about. But then, are you connecting with the person using the bot? Maybe, if there’s heavy curation involved.
If a writer uses a text-generator AI to create their writings, they’re not really a writer, they’re an AI curator. There’s nothing wrong with that — they’re just slightly different jobs.
It’s like when you see a picture of a beautiful place: do you love the place, the picture, or the photographer? As the husband of a photographer, I know photography isn’t just capturing, there’s a lot of creating involved — you don’t take a picture, you create a picture. And AI-generated text can be like that: you don’t just take what the bot spits out, you review it, curate it, and edit it until it reflects what you want to say. However, if the curator is just selecting what works or what sells — even if it doesn’t reflect their values and opinions — the reader who feels a connection with the writing can’t be connecting with the curator.
So if you’re not connecting with the bot, and you might not be connecting with the curator, who are you connecting with?
Connecting with yourself
Words are real-world magic. We take abstract things called ideas that live in the non-fungible universes of our minds and we encode them into symbols other humans can decode to recreate those same ideas in a whole new universe — their own minds. But when a bot writes a piece of text, only part of this process is happening. There is no origin for the encoded idea, no mind behind that thought.
The closest you can get to a source, in this case, is humanity: the source upon which the bot has been trained. Humans created the words and ideas you’re connecting with, so you’re connecting with humans talking about things that matter to humans. You’re connecting with the human experience, and so, ultimately, with yourself.
What’s next?
In a few years, we will see the first big movie made from an AI-generated script. I’m sure there is a bunch of such scripts already making their way to important hands in Hollywood, but they’re just not there yet.
Maybe we’ve already watched a movie with scenes written by a bot and didn’t even notice — why would big studios let us know?
When the first bot-written movie makes the news, I’m sure it will be all we talk about for a few weeks: everyone will want to watch it and see if it’s any good. When that happens, we’ll have more people trying to answer the questions I’m asking in this article.
Did I do that?
What I think will happen then — and is already happening — is humans will start asking themselves: “was this created by a human?” We will develop a sort of inner algorithm to recognize AI-created content.
Do you know how we’ve learned that a Nigerian prince doesn’t need our help, and that that girl you’ve met online is actually a dude in a dark room, and that we shouldn’t click shady links? It’s like that. We’re already training this algorithm when we see bots on Twitter, and our mental algorithm to tell humans and bots apart is relatively good, but as the bots’ algorithm improves, ours will have to improve too.
AI-scripted humans
I also think there will be “scandals” where people will discover that their favorite influencer is not actually the one behind their unique ideas. “Scripted by a bot” headlines will say. At that time, actual human ideas will become more valuable, just like hand-made products are sometimes more valuable than mass-produced ones. In a world of AI-generated fiction, humans might want to consume fiction created by other humans.
Is this a bad thing?
I don’t think of this in terms of good or bad. It will happen and we will have to deal with it. Perhaps having machines that can pass for humans will make us think harder about the question we’ve been asking our entire existence: what does it even mean to be a human?
In the meanwhile, if you’re one of the curators using AI to help you with ideas, I’d say the honest thing to do would be to label your articles as such. As a reader, I feel cheated when I connect with an idea, only to find the idea wasn’t created by the author. Let us know if your post, your tweet, or your script was created by a bot. It will normalize it for everyone.
F.J.C. Montenegro has a master's degree in artificial intelligence and robotics and loves writing fiction (and the occasional non-fiction).
