How the Ad World Shaped AI Writing-and Why Writers Should Care
AI writing tools have become ubiquitous in newsrooms and content departments. They promise efficiency, polish, and consistency. But they also homogenize voice, standardize tone, and strip away the friction that produces original thought.
The roots of this problem run deeper than recent AI breakthroughs. For decades, advertising has trained us to accept a particular kind of writing: upbeat, confident, ingratiating, commercial. AI language models have absorbed that aesthetic wholesale. When you ask them to improve a sentence, they don't just correct grammar-they apply a template.
The Template Problem
Test this yourself. Ask an AI to rewrite "We have lost our nose for truth." It will likely return something like "We have lost our instinct for truth." The change seems minor. A human editor might make the same call.
But when the same template applies to everyone's writing, endlessly repeated, those tiny alterations accumulate. Original phrasing disappears. Juice evaporates. What remains is serviceable, clean, and interchangeable.
A human editor works with one writer's voice. An AI trains on millions of documents and produces a voice that belongs to no one.
The Job Market Shift
Tech companies and AI startups are actively recruiting writers to train language models. Job postings for "AI trainers" have surged 150% over the past two years, according to Google data.
The work involves reviewing and annotating model responses-essentially teaching AI to sound more human. The irony is direct: writers are being paid to make machines indistinguishable from themselves.
Meanwhile, job descriptions for communications roles now read like advertising copy: "Creates effective communications content and plans focused on specific initiatives and changes." That language didn't emerge by accident. It's the voice of the marketplace, now baked into how companies describe what they want.
What Gets Lost in the Process
Writers often discover what they actually think through the writing process itself. Confusion, anxiety, the risk of being wrong-these are features, not bugs. They're where real thinking happens.
If the goal is a superficial, well-written piece, you can skip all that. AI excels at skipping it. But you also skip the moment when you find yourself "far out on some shaky limb" you didn't see coming-the moment when something worth saying emerges.
The problem compounds when writers internalize the advertising aesthetic as normal. Self-doubt becomes repackaged as self-improvement. We seek experts to tell us how to sleep, influencers to tell us how to look. That dependency doesn't serve clear thinking.
Control Through Language
Controlling language powerfully controls minds. AI writing, for all its convenience, represents a form of control that's rarely named as such.
The tool is seductive because it works. It produces lovely sentences without effort. But an invasive species can be beautiful-deadly lionfish wear colorful stripes to fool their prey. The question isn't whether AI writing looks good. It does. The question is what we lose when we stop doing the work ourselves.
AI has legitimate applications. But AI writing as a replacement for human expression is doing something else entirely: it's training us to accept a uniform voice as normal, to mistake polish for substance, and to value convenience over authenticity.
If writers don't reclaim sovereignty over their own language, we cede control over how we think and what we're capable of saying.
For writers working with AI tools: Consider AI for Writers resources that help you maintain control over your voice and process, and explore Prompt Engineering techniques that let you work with these tools without surrendering to them.
Your membership also unlocks:
