Lately, I’ve started noticing a pattern that’s hard to ignore. I open LinkedIn, and suddenly people who rarely used to post are sharing polished takes on technical topics. The structure is almost identical every time you look for it. A clean opening, a few balanced points, and a closing question designed to pull in engagement.

At first, I thought maybe people were just getting more active, more comfortable putting their ideas out there. That would have been a good thing. But the more I read, the more obvious it becomes. Not because the ideas are wrong, but because they don’t sound like the people writing them. They feel generated.

It’s the consistency that gives it away. The same rhythm, the same phrasing, the same type of conclusion that tries to wrap everything up neatly and hand it back to the reader. Once you start recognizing that pattern, it becomes hard to ignore. It stops feeling like someone thinking through a problem and starts feeling like something assembled to sound like thinking.

And that’s the part that bothers me. Not that people are using AI. I use it too, and I rely on it more than I probably should. But there’s a difference between using a tool to sharpen your thinking and using it to replace it. More and more, it feels like we’re drifting toward the second.

Part of me looks at this and thinks we’re heading in the wrong direction. Not because the content is wrong, but because it feels like the thinking behind it is getting thinner. It’s as if the goal has quietly shifted from understanding something to simply producing something that looks like understanding.

You can see it in how predictable it all becomes. The same structure, the same phrasing, the same attempt to sound balanced and thoughtful without actually taking a position. It gives the appearance of depth, but once you read closely, there’s not much there. It doesn’t challenge anything, it doesn’t commit to anything, and it rarely says something you couldn’t have guessed in advance.

And the more this pattern shows up, the harder it is to take any of it seriously. Not because the people behind it lack capability, but because the output no longer reflects their actual thinking. It reflects something else. Something optimized to sound right, rather than to be right.

That’s where the concern starts to grow for me. If the process becomes this easy, if producing something that looks thoughtful takes almost no effort, then what incentive is left to actually think things through?

Then again, it’s hard to blame the tool for how it’s being used. AI is doing exactly what it’s designed to do. It takes incomplete input and produces something coherent, structured, and readable. If that output feels generic, it’s because the input was generic to begin with.

In that sense, AI isn’t lowering the quality of thinking. It’s exposing it. It makes it obvious when there’s no real position behind the words, no real effort behind the idea. And maybe that’s uncomfortable, but it’s not necessarily a bad thing.

Because when it’s used properly, the difference is just as obvious. Give it clear direction, real constraints, and actual thought, and it becomes something entirely different. It stops sounding generic and starts sounding like you, just sharper, more structured, more deliberate.

So maybe the issue isn’t that AI is replacing thinking. Maybe it’s forcing a separation between those who think and those who just produce.

This shows up even more clearly in my day-to-day work. When I review traffic studies and send comments back, I expect a response that reflects actual engineering judgment. Not just a summary of what I said, but a position. A decision. Something that shows the problem was understood and evaluated.

What I often get instead looks polished, but doesn’t go anywhere. It acknowledges the comment, repeats parts of it, adds some general language about coordination or future consideration, and moves on. On the surface, it reads well. But when you step back, it hasn’t actually addressed the issue.

The problem isn’t that the response is wrong. It’s that it avoids committing to anything. It doesn’t take the data, the context, and the guidance we all know, and turn it into a clear recommendation. It stays in a safe zone, where everything is technically acceptable, but nothing is actually decided.

And that’s where it starts to break down. Because in this field, that’s the whole job. We’re supposed to look at the conditions, apply what we’ve learned, and make a call. Those calls aren’t always easy, and reasonable people can disagree, but they need to be grounded in something. When a judgment is supported and explained, even if it’s not perfect, it’s still something you can evaluate and respond to.

When that doesn’t happen, when the response reads like it was assembled to sound complete rather than to be complete, it becomes hard to trust what you’re reading. At that point, it’s no longer about disagreement. It’s about the absence of judgment.

And this is where the concern starts to extend beyond individual posts or responses.

If this way of working becomes the default, if more and more people rely on AI to generate content without putting in the thinking upfront, then the output will naturally start to converge. Not because the tools are identical, but because the inputs are. The same shallow prompts, the same surface-level intent, producing the same kind of structured, reasonable-sounding responses.

Over time, that starts to compound. That content becomes part of what’s out there. It gets read, shared, and eventually learned from again. And if the cycle continues, it’s hard not to wonder whether we’re slowly moving toward a kind of equilibrium where everything starts to sound the same. Different people, different contexts, but increasingly similar expressions of thought.

Part of me thinks that might be overstated. That people will adapt, that real thinking will always find a way to stand out. But another part of me isn’t so sure. Because the easier it becomes to produce something that looks thoughtful, the less pressure there is to actually be thoughtful.

And that’s the part I can’t quite shake. Not that AI will replace thinking outright, but that it might quietly make it optional. That it might blur the line between something worked through and something generated to the point where it becomes harder to tell the difference, even for ourselves.

If that happens, then the question isn’t just about authenticity anymore. It’s about whether the way we think, and how we communicate that thinking, starts to change in ways we don’t fully notice until we’re already there.

And if that line keeps blurring, it’s worth asking how much of what we’re saying is actually ours.

.

.

. P.S. Here’s an example: I used AI as an editor on this piece, not a substitute for thinking.