Being an ageing content writer may have made me comfortable with AI

My comfort with AI might say more about where I am in a career than about AI.

Being an ageing content writer may have made me unusually comfortable with AI.

I’ve been trying to work out why — and I’ve come to think the comfort is worth examining rather than trusting, because it arrived far too conveniently.

I’ve written professionally for nearly twenty years. Journalism, copywriting, B2B content, interviews, customer stories, SEO, thought leadership, campaigns. I know I can write. I’ve produced enough 1,500-word articles from blank pages to last several lifetimes.

So when AI makes the first version of those sentences much easier, I don’t feel that something fundamental has been taken away from me.

The career I chose probably explains more than the technology does

I never really followed the obvious route into middle management.

I stayed unusually close to the work: interviewing people, reading research, finding angles, understanding unfamiliar subjects and trying to work out what was actually interesting about them.

Over time I became less interested in writing itself and more interested in what happens before it.

What do we know?

What’s missing?

What’s the argument nobody has made yet?

Which evidence changes the story?

Why is everyone repeating the same thing?

Should this piece exist at all?

That’s probably why AI clicked for me so quickly. It doesn’t feel like it’s replacing the most valuable part of what I do. It gives me more leverage over the part I’ve spent twenty years learning.

Which is a very comfortable thing to believe, and I should say plainly that it’s also extremely convenient. A person whose main skill was producing sentences quickly would look at the same technology and reach a different conclusion, and they wouldn’t be wrong to.

The part I never liked is the part that’s gone

I should be specific about what I was reacting against, because it started long before AI.

As a journalist I was frequently frustrated by how much of the job was cut and paste — from technology company announcements, from other people’s reports, from whatever the industry had already agreed was the story that week. B2B can be worse. There are corners of it where a piece is assembled almost entirely out of material that already exists, rearranged into a slightly different order and published as though something had been added.

I always gravitated towards trying to find something new. A different angle, an argument nobody had made, a piece of evidence that changed the shape of the thing.

And that was always the hard part. Not the writing.

What do I write that is new and hasn’t already been said?

That question is the one I’ve never found a shortcut for, and rehashing arguments or working to the numbers was never a version of the job I was comfortable doing.

Which brings me to what I think is actually being mourned.

The complaint is usually described as being about writing. Much of it is about the act of writing — the craft of assembling a competent piece to a brief, at speed, to a reliable standard. That is a real skill, it took years to acquire, and for a great many people it was the job. Watching it become cheap is a genuine loss and I’m not going to be breezy about it.

It just was never where my satisfaction lived.

Which is probably why AI turned up looking like a tool rather than a threat. It doesn’t help me with the question. It gives me a way to pursue it at a pace a business can actually absorb.

I’ve watched a definition of value collapse before

The other thing that helps is a long memory of digital publishing.

I lived through the SEO era, when a piece could carry more organisational value because it ranked for a keyword than because it contained an original idea. Enormous quantities of derivative content were written by people, slowly, because search traffic justified the expense. A lot of it was still slop.

I’ve made that argument in full elsewhere, so I’ll leave it at the part that matters here: I have already watched a production-shaped definition of editorial value collapse once. That makes me harder to alarm when another one starts to wobble, and it’s why I don’t experience this as an ending.

For B2B work, I increasingly think the valuable part was always judgement, evidence, synthesis, customer understanding and originality of angle. Not the physical act of typing every sentence.

So I’m comfortable giving AI a rough thought, letting it work across the context I’ve accumulated, seeing the argument reflected back, disagreeing with it, finding what’s missing and editing from there.

I don’t use AI because I can’t write. I use it because I don’t have unlimited hours and I have more ideas than I can manually turn into finished pieces.

Why the upstream work is only now practical

Something follows from this, and it’s the reason I’ve spent a year building a methodology rather than simply holding opinions about one.

The work I care about sits upstream. What do we know? What’s missing? What does the evidence actually support? Should this exist at all? I’ve argued separately that editorial jobs are moving in that direction, so I won’t make the case again here.

What I’d add is why it didn’t happen sooner.

Upstream work was always the valuable part. It was also, until very recently, slow enough to be impractical. Reading everything, connecting it, establishing what has already been said, finding the actual gap, then testing whether an angle survives contact with the evidence — that is weeks of work for an argument that may not hold up. Most organisations could not afford to find out, so they skipped it and briefed the writing instead.

Which is how those questions came to be passed around rather than answered. To the agency, to the brief, to the stakeholder, to whoever ended up writing it. Everybody assumed somebody upstream had already done the thinking, and frequently nobody had.

Editorial Intelligence is mostly an insistence on doing that work rather than handing it on.

AI didn’t make it valuable. It made it affordable.

And that is a strange sort of luck. The part of the job I was drawn to for reasons that had nothing to do with strategy is the part that has just become possible to do properly, at a pace a business will tolerate.

Where my position stops working

This is where the apparently pro-AI position gets messy, and I’d rather show the mess than tidy it.

I don’t feel remotely as certain about music. Or films. Or novels. Or illustration.

There, the human act of making the thing feels much more entangled with why I value it. Somebody lived something. Developed a voice. Learned an instrument. Made choices. Failed. Tried again. Expressed something.

That feels different to producing a useful B2B article about accounting software. I think it genuinely is different.

And then I hear an AI-generated song I enjoy.

Which is inconvenient. Because if I enjoy it, I enjoy it. I can’t decide afterwards that the emotional response wasn’t real because I dislike what the technology might mean for musicians.

Here is the uncomfortable version, which I’ve only arrived at by writing this down.

My boundary may not be a principle at all. It may just be a description of which work I know well enough to see past the production, and which work I don’t.

I understand B2B writing from the inside, so I can see that the typing was never the valuable part. I don’t understand how music is made, so the performance and the value look like the same thing to me. A musician might look at my B2B argument and see somebody casually dissolving the craft in their own field while defending it in theirs.

I can’t rule that out. It’s the strongest objection to everything I’ve written about this, and it comes from noticing that my confidence maps suspiciously neatly onto my expertise.

What I’m least sure about

Two things, and they’re both about who this position is available to.

It may be a function of having already done the twenty years. Everything I’ve described depends on having accumulated judgement, context and enough finished work to know what I’m capable of. Somebody four years into their career has the same tools and none of the standing, which is a problem I’ve written about separately and haven’t solved. My comfort is not transferable advice.

I can’t rule out that I’m rationalising relief. I have spent two decades doing something that is now much easier. Constructing an account in which the part I’m good at was the valuable part all along is exactly what somebody in my position would do, whether or not it’s true.

Writing that down doesn’t make it false. It just means the argument has to stand on the evidence rather than on how satisfying I find it.

The position I’m most comfortable with

Not being pro-AI or anti-AI.

Trying to work out where human value actually sits as the tools change, and accepting that the answer is different in different places, that the boundaries between those places are less tidy than I’d like, and that I’m on one side of several of them for reasons that have as much to do with what I know as with what’s true.

After nearly twenty years in content, I’ve seen enough supposedly permanent rules disappear to be suspicious of simple answers.

Including my own.