AI didn't create generic content. It exposed it

A strange thing happens when people complain about AI slop on LinkedIn. The posts making the complaint often look remarkably similar to the posts they’re criticising — short paragraphs, a simple opening, a clear opposition, a conclusion that can be understood without much effort. Sometimes they may even have been written with AI. But if the argument connects with people, the post performs well: readers agree with it, add their own experiences and share it with others.

Apparently AI-style writing is not always the problem. We object to it when we dislike what it’s saying, when it feels empty, or when we can see the author hasn’t done much thinking. When the same accessible structure carries an argument we recognise, we’re much more forgiving — perhaps because humans like ideas presented simply.

That’s not necessarily a weakness. Most people don’t open LinkedIn wanting to decode difficult prose; they want to understand something quickly, and familiar structures reduce the effort required to follow an argument. The problem begins when simple writing gets confused with simple thinking.

Generic content existed before generative AI

Content writers have always worked under pressure. Freelancers had deadlines, word counts and fees that rarely allowed for days of original research. In-house writers had publishing calendars, campaign requests, stakeholder reviews and targets to meet. Usually we were given enough time to write the article. We weren’t always given enough time to work out whether it was the right article.

So we searched online, opened the highest-ranking results and assembled something from the available material. We followed a proven format, added examples, filled out the required length and moved on to the next assignment. The result was often a competent but interchangeable article:

“Seven ways to improve productivity.”

“Five trends changing the future of work.”

“Ten reasons your business needs a new accounting system.”

Much of it was written manually. That didn’t automatically make it original, useful or good.

Writers weren’t necessarily being lazy — often this was what the system allowed. Businesses wanted a dependable stream of searchable content. Agencies needed to protect their margins. Writers had to meet deadlines. Search strategies rewarded familiar questions, predictable structures and comprehensive-looking lists.

Generative AI didn’t invent this content model. It learned from it. That may be why so much AI-generated content feels familiar: it resembles the internet we spent years producing.

AI has removed part of the production excuse

AI can now help with desk research, transcription, note organisation, outlining, drafting, editing and repurposing. Tasks that once consumed several hours can sometimes be completed much faster.

That doesn’t mean an AI-assisted article is automatically better — the easiest response is to use the saved time to produce even more generic content. But that’s a choice a business makes, not something the technology forces on it.

The more interesting option is to reinvest the time. A writer can dig into the subject more deeply, find stronger evidence, talk to somebody with direct experience, test the central claim and work out what the existing coverage missed.

They can ask whether the article contains a real argument or just reorganises information that’s already out there. They can spend longer editing it — cutting the passages that only exist to hit a word count, making the piece clearer without flattening its ideas.

That’s not abstract for me. For years, my head was down on content — hit the brief, meet the deadline, move on to the next piece before there was time to ask whether the last one actually said anything. I worked hard the whole way through. But the only evidence of that effort was more content: drafts that still needed editing, still needed another pass, still needed time I didn’t have left to give them. Working hard and having something to show for it were not the same thing.

AI has made producing plausible copy easier. That means plausible copy is no longer enough — the value has to move upstream into the thinking.

The writer’s job should not end at publication

There’s another opportunity in the time AI saves — the chance to actually think about a piece before it goes out. Not just whether it’s finished, but whether it’s saying anything, and what happens to it once it’s published.

Traditionally, most writers delivered the article and moved on. Promotion, distribution and measurement belonged to somebody else — if anybody owned them at all. That division made sense when writing the article ate up most of the available time. It makes less sense now.

Writers can spend more time helping the work surface. They can turn the central argument into a series of social posts, rather than just copying excerpts out of it. They can connect the article to related research, customer stories and commercial pages.

They can help sales teams understand when it’s actually useful to send something to a customer. They can adapt it for newsletters, presentations, video or AI search. And they can look at which parts actually got a response, and use that to make the next piece better.

Most importantly, they can return to the original purpose. What was this article supposed to do — change how a particular audience understands a problem, support a product decision, establish expertise, generate conversations, give existing customers something genuinely useful? Publishing isn’t the outcome. It’s the point at which the article gets the opportunity to produce one.

AI could allow writers to take more responsibility for that entire journey. The job becomes less about supplying words and more about helping an idea achieve something for the business — harder to measure than the number of articles published, and potentially much more valuable.

But do readers actually want better content?

This is where the argument gets less comfortable. Writers often assume people are waiting for greater originality, complexity and depth. But audiences keep choosing familiar subjects, recognisable opinions and simple explanations.

A generic-looking post with a clear argument can outperform a deeply researched article. A list can be useful. A straightforward summary can be exactly what somebody needs.

We should be careful not to make “depth” another way of flattering ourselves. Readers don’t owe writers attention just because something took a long time to produce. An article can be original and still be confusing, self-important or irrelevant — complexity isn’t proof of quality.

A lot of writers’ instinctive defence of AI-generated content is really a defence of effort — and effort was never actually what readers were paying for.

Perhaps the real distinction isn’t between simple and sophisticated content. It’s between content that makes a decision and content that avoids making one. A simple article can decide which evidence matters, what the reader should understand and why the subject deserves attention. A long, heavily researched article can still hide behind summary and never actually reach a conclusion.

That might also explain why some supposedly AI-generated posts perform well. The structure might be predictable, but the argument gives people something to agree with, reject or add to. The author has made a decision.

What readers seem less willing to tolerate is content that feels both familiar and empty.

The end of the listicle may not be a tragedy

Generic listicles once served several purposes: they filled publishing calendars, targeted search terms and gave businesses something regular to distribute. AI can now produce them in seconds, which makes their lack of distinctiveness impossible to ignore. That may be bad news for a particular model of content production. It isn’t necessarily bad news for writing.

If a business can no longer justify publishing another generic list of tips, it may have to offer something that can’t be assembled from the first page of search results — original evidence, a useful tool, a strong opinion, direct experience or a connection that hasn’t already been repeated hundreds of times. Writers may produce fewer articles, but they can spend more time making each one distinctive, helping it reach the right audience and learning whether it worked.

There will still be formulaic content, because it’s cheap, easy and sometimes effective. AI won’t automatically make the internet better written — it might just flood it with more of the same.

But it has exposed how much human-written content was already produced without enough research, judgement or purpose. That leaves a choice: use AI to speed up the old content model until readers stop paying attention altogether, or accept that producing words is no longer the most valuable part of the job.

The future role of the writer may begin before the brief and continue long after publication — deciding what’s worth saying, developing the argument, making it accessible, helping it surface and establishing whether it made any difference. That world demands more from writers than the SEO-listicle era did. It could also be considerably better.