AI slop.

That's a term we're hearing more and more. I don't remember hearing the word "slop" all that often before. Now it's sneaking in everywhere as more and more AI generated content makes its way to the internet.

Slop is what pigs eat. A nice connotation isn't it?

I've been grappling with how much to use generative AI in my workflow for years now. And one of the most obvious and common uses is for writing.

I've written before about how I entered a contest that forbade using AI in the drafting process. It felt like cheating and it was cheating for that contest because it specifically had rules against it.

But I've been exploring this more and more as I've been working on my own creative writing projects.

I've had these ideas for storyworlds that I've been wanting to develop for years and I've been making excuses not to write them. I've found a lot of reasons to disqualify myself. I was suffering from a form of what many call writer's block.

There are two forms of writer's block that are hard to tell apart. Sometimes it's a deep psychological protective mechanism around avoidance, fears, and unresolved traumas that need to be addressed. Other times it's like a cold engine that needs a little kickstart. LLMs gladly provide the kickstart and may mask the psychological needs.

These two forms feel almost identical and it requires honesty to navigate.

I've been confronting my own psychological blockages around writing. I've unearthed a lot of fears around failure and criticism that stem from early memories. I've also found a lot of ways to say I'm not qualified.

In the writing process, I've been trying large language models a few different ways. I've tried using it to interrogate me and pull ideas out of me — kind of like Socratic dialogue. I feel the best about that one because I know it's truly helping get thoughts out of my brain. I also use it for collaborative brainstorming where it helps me add details to my architecture. I've had it create an outline that I then use to prompt me to write out full scenes. The one that gets the trickiest is generating first draft that I then rewrite.

The first feels like it's the least risky of me losing my voice and creative ownership. And then it gets progressively riskier when you think about "ownership."

I started slipping into having a large language model generate drafts for me that I rewrote and typed out letter by letter. I did not copy what it wrote. I typed it out myself.

But something about that didn't feel right. I felt guilty about it.

I went through a deep reflection process and realized that I'd turned using LLM assistance into the latest thing to make me feel not qualified about my writing.

If I took that vague "qualification" requirement out of it, I could look at my work and feel proud of it. I was having FUN. My ideas were coming to life. Finally.

There's never been a stable definition of a "real" author. It has always been gatekept.

Real authors get agents. Real authors are traditionally published. Real authors don’t write genre fiction. Real authors don't use outlines. Real authors don't use dictation tools. Real authors only use typewriters. Real authors only handwrite. Real authors don't have ghost writers. Real authors don't self-publish.

And so on! The list could go on forever. All of these have been gates that have held writers back from genuine creativity.

LLMs are the newest version of that gate. Instead of me patterning myself around modern fiction standards which have become extremely formulaic, I can focus on the vision and then deliver it the way I want it to be without worrying about reconditioning my brain to write commercially.

I've learned in the process of experimenting with LLMs that it takes a significant amount of time to get it to truly capture my vision. Days of worldbuilding, plot refinement, and back and forth before a single scene gets drafted.

The important question I always consider: Where is my voice?

That's going to look different for everyone. What I decided is that, for me, it's more important for me to keep my voice in the essays, blogs, and nonfiction books that I write. That IS me. That is me articulating my ideas. And that's been incredibly healing. More healing than any course I've taken.

Maybe I am losing something in the process of not crafting these minute character or sensory details in a fictional world. Or maybe it's freeing me up to create my other content I feel called to create.

It's hard to know. We just have to make a choice.

I feel that I've been born in this era with the ideas that I have to leverage the assistance of the modern tools that I have today. And one of those tools are large language models. Other tools include web hosting, laptops, smartphones, word processors, spell check, and Google.

Every generation of writers had something that previous writers might have called cheating or illegitimate.

The question that persists through the name calling is: "Did something true, human, creative, and meaningful come through?"

Ultimately, the question is not about the tool. It's about whether you remain the one who knows what the work is for, what it means, whether it's true.

I'll be transparent. I haven’t always gotten this right. There's some older content on my blog that was written by hired writers who used AI without my knowledge or discernment at the time to catch it. That's partially now how I know what AI slop sounds like.

And why I take the question of my voice and authenticity seriously now.

Moving forward, it is my aim to be able to answer that what I create answers that question with a resounding yes.

Navigating this with you,

Rachel

P.S. — The framework I used to find my own footing with all of this lives here: [The Zones of Technology →]​

P.P.S. — Want to apply it specifically to AI? I made a guide for that too: [The Zones of AI →]​