Transform, Don't Generate
Stop generating slop. The writers who won't get caught use AI to transform their ideas, not replace them. Here's the difference.
A recent Forbes article caught our attention with its clickbait title: 6 Giveaway Signs Of ChatGPT-Generated Content.
Among the tells are:
Lengthy introductions ("throat clearing") - Rambling opening paragraphs with extraneous backstories or metaphors before getting to the main point
Overuse of ethical considerations - Tendency to constantly cover all bases and carefully weigh both sides of any argument
Generic, noncommittal thoughts and advice - Failure to take a strong stance due to the model's training on conflicting viewpoints
Lack of personal stories or examples - Inability to draw from personal experiences or observations like a human author would
Employment of specific overused phrases - Reliance on favorite vague chatter like "unleash" or "buckle up" as conversational filler
Formulaic article structure - Rigid pattern of intro, 3-4 body sections, further considerations, summary, and ethical wrap-up
The Article is Informative on Two Levels
First, it shows you what to watch out for if you're trying to use AI to write without getting caught.
But on a more fundamental level, it reveals a common assumption that the best (only?) way to use AI in a writing capacity is to have it draft the words for you — generating a coherent, but often vague and platitudinous word soup.
So much sound and fury, signifying nothing.
The advice in the Forbes article is golden for anyone relying heavily on basic AI generation prompts — things like "write me a listicle about how to make ChatGPT sound less robotic." However, more advanced users have discovered even greater value in content refinement. Specifically, transforming your own raw materials and ideas using AI assistance. This approach avoids the common tropes that betray AI authorship altogether because you are performing a different operation: transformation, not generation.
AI Excels at Enhancement
AI is so-so at conjuring up full compositions, but it excels at enhancing existing content.
Tools like Notion and Grammarly take advantage of this strength by building in editing features like the "Improve writing" prompt to streamline writing workflows within the documents where you do your writing.
I first discovered the power of AI when I used Notion AI's improve writing prompt to save me hours of revisions while polishing transcripts for podcasts.
In this case, an AI generated transcript serves as a kind of rough draft, or brain dump. You can polish it, while leaving the basic structure of dialogue intact, or you can use it as the "Source material" for an outline — weaving the best-worded sections together into a tight-knit composition, from the meandering conversation.
Transforming transcription as source material overcomes the dual problems of generic content and lack of personal stories or expertise. At least, as long as your source material contains something worth saying.
The "Improve Writing" Superpower
The "improve writing" prompt in Notion may be the single most valuable AI feature for writers, allowing them to polish their rough ideas into shining gems with just a highlight and a quick keystroke (Command + J). This ability to transform text within your workspace, as opposed to the chat interface, allows you to stay in the flow while you write and revise.
But Improve writing is just the tip of a much larger iceberg of "transformation operations" you can use to restructure or otherwise enhance your own source material. Most people are aware of AI's translation and summarizing capabilities. If AI is a magic wand, think of each of these operations or commands as spells — incantations that form a certain kind of prompt for turning this into that.
Of course, if your source material contains little of value, there's not much AI can do. It's a bit like adding an Instagram filter to a crappy photo. Sure, ChatGPT can transform the worst writing into iambic pentameter in the style of William Shakespeare. But once the novelty wears off, your readers will see through the ruse.
Framed differently: Garbage in, garbage out.
The flipside of this old adage is treasure in, treasure out.
The Fairy Godmother in Cinderella worked her magic because of the underlying substance of the material she was working with. Cinderella, though unpolished in her tattered clothes, had a heart of gold. So did her friends.
So before you start transforming content, you'll need a way to capture your rough "ore" — the thoughts and ideas through brainstorms, voice memos, stream-of-consciousness writing sprints, etc. These activities generate the raw materials for AI to refine into sparkling gems.
Create an AI Content Pipeline
This is how I've been using AI to repurpose and transform ideas:
Record a podcast or capture thoughts via voice memo
Use a service like Otter.ai or Descript to transcribe the audio
Load transcripts into a workspace like Notion and use Improve writing to create a draft
Redistribute polished excerpts as social media posts, newsletter updates, etc.
By understanding the principles of prompt engineering and building customized workflows that leverage AI's strengths in content transformation, you can unlock new levels of productivity and creativity.
Your unique perspective, expertise, and personal stories will always be the secret ingredient that sets your content apart from the mass of AI-generated pablum.
Embrace AI as an ally, but don't forget who's the master.