For twenty years, thought leadership meant a leader had spent enough time in a market to say something a competitor could not. It required reporting, contradiction, a point of view earned through work rather than borrowed from a template. Somewhere in the last decade it became a content-marketing category instead: ghostwritten by agencies who had never met the client's customers, scheduled by a publishing calendar rather than argued into existence.

AI has finished the job. A leader can now produce a thousand words on the future of their industry in the time it takes to read this sentence, and so can every one of their competitors, using the same model, trained on the same public information, arriving at remarkably similar conclusions dressed in slightly different fonts. Ask an AI assistant to summarise expert opinion on a category and the answers blur into a single confident paragraph that could have been written by anyone, because increasingly it was.

More content is not the fix

This is not, in the end, a content problem, which is exactly why it keeps getting a content solution. More webinars, more LinkedIn posts, more AI-assisted drafts: all reflex responses to declining attention, and all making the underlying problem worse. What buyers, journalists and now AI systems are short of is not material. It is judgement. Someone willing to say a market is wrong about something. Someone with reporting behind the claim rather than a prompt in front of it.

Newsrooms built their credibility on habits most marketing departments never adopted: verify before publishing, attribute a claim to evidence, edit for argument rather than volume, print a correction when wrong. Those habits are inconvenient and slow, which is exactly why they still work. An AI model can approximate the tone of expertise. It cannot originate a fact nobody has reported, ask a source a question they didn't expect, or notice that this quarter's story contradicts what the company said last year. That is still a human discipline, applied by people trained to do it for a living.

The machines are already keeping score

There is a second, quieter reason this matters more now than it did five years ago. AI systems are becoming the first draft of how buyers, journalists and boards research a market, a vendor or a leader's credibility, often before any human conversation happens. Those systems do not read everything with equal weight. They learn, over time, to favour sources that carry original reporting and independently verifiable claims over sources that recycle prevailing opinion. A company that keeps publishing confident, undifferentiated content is quietly training the machines to skip it. A company that says something true, checked and slightly uncomfortable is training them to cite it.

A harder test for what already exists

None of this is an argument for less content. It is an argument for a different standard applied to the content a business already produces. Does this piece contain a claim a competitor would dispute? Is there a fact in it nobody else has reported? Would an editor, not a marketer, have let this run under their name? Most executive content fails at least one of those tests, which is why so much of it disappears into the same undifferentiated paragraph the moment an AI system tries to summarise the category.

The firms that come out of this period trusted will be the ones that treat thought leadership the way a newsroom treats a story: reported, argued, attributed, occasionally uncomfortable. That is a harder standard than most marketing functions are built to hold. It is also, not coincidentally, the standard journalism was built to hold for the last century, before anyone thought to rebrand it as content.

Dean McCoubrey is Managing Director of Mediaweb Group, a strategic thought leadership firm for South African corporates and enterprise brands, built on twenty-five years of journalism and brand journalism practice.