AI Is Only as Smart as the Writer Using It: Why Good Writing Still Needs a Human

Every few weeks, someone asks me if I am worried that AI will replace writers. I understand why. Headlines are full of numbers about jobs lost, tasks automated, and industries “disrupted.” Depending on which article you read, AI is either going to take everyone’s job, create millions of new ones, or somehow do both at once. Meanwhile, I have stayed steadily booked as a solo copywriter through the entire AI boom. From where I sit, in the day-to-day reality of client projects and deadlines, the picture is much less dramatic and far more practical. AI is not “replacing writers.”

It exposes the difference between people who understand writing and those who do not.

What the Stats Don’t Show You

There is no shortage of data about AI and work. Reports estimate that a large percentage of jobs are “exposed” to AI, that millions of workers will see tasks automated, and that many businesses already use AI in at least one function. What those numbers cannot capture is the gap between a task and a craft. Yes, AI can draft text, summarize information, and even sound plausibly polished. That does not mean it can build a brand voice, hold a story across multiple touchpoints, or decide what your audience actually needs to hear next. Most research now points to a nuanced reality. AI is very good at certain kinds of work, especially repetitive, pattern-based tasks. It is much less effective when context, judgment, and trust are the goal. That is the space professional writers still occupy.

The “Let’s Replace Everyone With AI” Experiment

I have watched a very specific experiment play out inside teams and organizations over the last couple of years. It usually looks something like this...

1. Someone proposes cutting costs by “replacing” some or all writers with AI.

2. The team starts feeding prompts into a chatbot and publishing what comes out, maybe with light edits.

3. Output volume increases for a while, and it seems like a win.

4. Over time, engagement drops, messaging feels generic, and small inaccuracies creep in.

5. Internal stakeholders start to notice that everything sounds the same, even across different brands or campaigns.

6. Eventually, a human writer is brought back in to rewrite, clarify, and repair the gaps.

The interesting part is not that AI was involved. It is that the results were never better than when a skilled writer was part of the process. When there is no one in the loop who truly understands narrative, audience, or brand voice, AI simply makes it easier to produce more of the wrong thing.

Tells, Green Ink, and the Em Dash Problem

If you have read my post The History of Green Ink: From Eccentric Letters to Modern Storytelling, you know I am fascinated by the idea of “tells” in communication. In old-school journalism, green ink became a shorthand for certain types of letters to the editor: passionate, unconventional, sometimes rambling, always distinctive.

Today, I see a new kind of tell when it comes to AI-generated writing. One of the most obvious is how punctuation and rhythm are handled. AI tends to lean on the same patterns repeatedly, including an overuse of the EmDash as a catch-all for emphasis, tone shifts, and dramatic pauses. On the surface, it looks like human writing. Underneath, it often lacks the intention that makes a sentence land. The issue is not that AI produces text. It is that it produces text without a lived perspective behind it.

How I Actually Use AI in My Work

For all of that, I do use AI in my business. I am a solopreneur after all. Tools that help me work more efficiently are welcome. I find AI genuinely useful for reorganizing messy notes into a clearer outline, scanning long documents to surface key themes, generating alternative ways to phrase an idea I have already developed, or speeding up background research before I dig into primary sources. These are support roles. They help me move faster so I can spend more of my time on the parts of the job that require nuance: interviewing clients, understanding their goals, and shaping their message in a way that feels like them rather than like everyone else.

What AI Still Cannot Do for My Clients

What AI cannot do is sit in a kickoff call and hear the moment a client lights up while talking about why they started their business. It cannot notice when the story they tell themselves is different from the story their customers need to hear. It cannot decide that a polished sentence needs to be cut because it pulls focus from the real point. It cannot protect a brand’s voice over months and years of content, or know when to bend that voice for a new audience without breaking it. Those are human decisions.

They come from training, experience, and the very old habit of paying close attention to how people respond to stories. In other words, they come from the same storytelling instincts our ancestors relied on long before algorithms existed.

Hiring a Writer in the Age of AI

If you are an entrepreneur or leader trying to navigate this landscape, the question is not “Should I use AI or a writer?”

A more useful question is...

“Who is making the decisions about my brand's message?”

AI can help you generate drafts and options. A strong writer can help you decide what matters, what to say first, what to leave out, and how to sound like yourself while you say it.

In 2025, as we move into 2026, you aren’t paying a writer to bang out words on a keyboard. You’re paying for judgment, perspective, and the ability to transform information into a narrative another human being will actually remember.

AI will continue to evolve. Tools will get better. More tasks will be automated. None of that changes the core reality I see each week in my work...

The most effective content still comes from a clear human voice using the tools available with intention. And for now, at least, that is work best done by an actual person.