With AI, we’re entering the era of language inflation — too many words are chasing too few ideas.
A curious thing has happened to writing.
As most writers know, words worth writing once arrived with a certain reluctance. Drafts were coaxed into existence, sentence by sentence, haltingly, painfully. Writers struggled to decide not only how something should be said but whether it deserved to be said at all.
The friction was real (and maddening), yet it served a purpose: the effort required to produce language tended to slow the mind just enough to let ideas take shape.
That friction has largely disappeared. Today a paragraph can appear almost instantly, summoned with a prompt and delivered in the composed, confident tone of professional prose. It’s like turning on a faucet that no longer sputters or resists but runs at full pressure from the very first second. And the results are subtle but profound.
Writing, which once required time and intention, has entered an era of near-limitless abundance. Words now circulate through organizations the way weather moves through an open landscape, arriving continuously, facing little resistance. And something odd has accompanied this abundance. As language becomes easier to produce, the difference between writing that fills space and writing that actually carries an idea has grown harder to see.
When Words Become Cheap, Judgment Becomes Expensive
For centuries, the first draft was where it all started. Writers carved into that rough material like sculptors into clay, discovering what they meant as they added and subtracted words. This stage has largely been automated now. Machines can generate paragraphs that resemble finished text, filled with the reassuring cadence of ‘good writing.’ But the surface competence of these drafts conceals a deeper problem, a subtle danger.
Readers begin to confuse fluency with insight, assuming that just because a paragraph reads smoothly, it must also be saying something meaningful. Yet writing has never been valuable simply because it has lots of words that sound good. Its value lies in the thinking that precedes the sentences and the judgment that shapes them afterward.
We are now entering a time of language inflation, even hyper-inflation. If inflation is defined as too many dollars chasing too few goods, there are, today, simply too many words chasing too few ideas.
In such an environment, the critical skill is no longer generating language but deciding which language deserves attention. Good writing now depends on the ability to recognize when an argument is unclear, when an idea has been diluted by too many qualifications, when a sentence imitates meaning but doesn’t actually deliver it. It calls for the discipline to ask whether a thought is worth expressing at all.
As the cost of producing drafts approaches zero, the craft of writing has to move elsewhere. We need less composition, more discernment. And this kind of judgment can’t be automated easily because it relies on context and experience. At the risk of sounding self-serving (my title is, after all, Editorial Director), the future of writing may belong less to those who produce the most words and more to those who know which words to keep — the editors.
Editors recognize the difference between language that fills space and language that carries an idea forward. They notice when a paragraph gestures toward an insight but never quite arrives, or when a sentence contains two competing thoughts that weaken each other. And at their best, editors do more than repair language; they introduce the idea that was missing in the first place, drawing together fragments of thought until a clearer argument begins to emerge.
Voice Becomes the Last Form of Scarcity
Generative systems draw from the accumulated language of the internet. This means their output will inevitably settle into patterns that are already familiar to readers. We get competent prose that rarely surprises. It moves efficiently from premise to conclusion while avoiding the idiosyncrasies, the peculiarities that make one writer, one message, one organization distinct from any other. Texts start to resemble the muddy brown we got as kids when we combined all the crayons — disappointing and kind of unnerving.
Of course, the tools that automate drafting are also incredibly useful. They remove many of the mechanical obstacles that once slowed the process of communication. But when we rely too heavily on AI, we lose the subtle qualities that make communication memorable: a point of view that’s shaped by lived experience, a rhythm of thought that reflects how real people speak, a willingness to say something unique rather than something acceptable.
What we call voice emerges from those choices. It’s not a stylistic flourish, it doesn’t come easy and it can’t really be faked. What AI-generated language is revealing to those who read carefully is that the real craft of writing has never resided in the words. And looking ahead, really good writing won’t be defined by the ability to produce fast, sleek text — machines have clearly mastered that task. Good writing will instead recognized by what remains stubbornly human: clarity about what matters, restraint about what doesn’t, and the patience to refine an idea until it finally captures a thought worth sharing with another person.