Lately I’ve been seeing a striking number of posts and articles where, a few lines in, it’s obvious: this is AI. And the ones I don’t spot? There are probably far more of those. Studies show we are much worse at recognizing machine-written text than we think.
And honestly, that isn’t a problem in itself. If you look at reach, likes and comments, plenty of it looks like success.
Still, one question won’t leave me alone. One that is more uncomfortable than “How do I write better prompts?”
When output costs nothing, output is no longer an advantage
AI lowers the cost of content. Paid media turns reach into something you can buy. Both can look very good in the short term. Neither answers the question of demand without pressure.
When the price of output approaches zero, output stops being a differentiator. That is also why “AI content” doesn’t automatically fail. It is fast. It is good enough. It feeds the algorithm. But this is where the problem begins: signals are not automatically substance.
I once produced this lesson myself, over two years. After my radio years, long before AI was a topic, I made a corporate podcast for a large brand. Solid craft, regular episodes, real effort behind it. For all that work, the result was modest: a few hundred subscribers. Lots of output, visible activity, and at the end the question nobody likes to ask: What did it actually achieve? That question has stayed with me. Only now, the output that took weeks of work back then costs next to nothing.
Engagement can be a false friend
With AI you can build yourself a machine that turns out material every day. You can even build momentum with it. The question is what it carries into the market. Trust? Preference? Demand that shows up even when nobody is pushing? Or mostly clicks and the next campaign?
Paid reach is a powerful tool. But tools turn dangerous once they become the business model. Advertising effectiveness research has been saying it for years: if all you ever do is activate, you wear the brand out. So here is the uncomfortable question any company can ask itself: Would the revenue still come if the budgets went down?
Why I’m not building myself a content machine
That sounds counterintuitive, especially now. But at the moment I’m not building an automated content machine for my own presence. I don’t draw the line at AI; I use these tools gladly and openly. I draw it at judgment: nothing goes out unless I have decided that this exact thing should be said. A machine that publishes every day would take that decision away from me, one piece at a time.
Because a system that publishes every day shapes your positioning over time. Once the machine is running, you quickly start optimizing for what works. And at some point you are known mostly for whatever the algorithm paid back the most. That is my observation. It is no law of nature. But I have seen it happen too often to ignore it.
And to be honest, I’m not selling a product that needs a daily supply of content. That is exactly why my bar sits somewhere else. Not “what brings reach?” but “what builds substance, and fits the stance I want to stand for?”
What can’t be generated
Real thought leadership has side effects. It isn’t for everyone. It often polarizes. And it forces you to be clear.
AI can help with the wording. It can’t help with the core: the decision to take a position. And in the end, that core decides whether content turns into demand. Or trust. Or just reach.
If a company cut its paid reach in half tomorrow, would it only see less reach? Or would part of the business disappear? And if so: What would have to start today for the answer to be different twelve months from now?