In the World of Artificial Intelligence, Be More Like Ted Lasso

I had one of those moments this weekend where you think you have something completely figured out, and then it unravels in about five minutes.

I was dealing with a situation between my wife and me, and instead of just talking it through, I decided to run it through Claude first. I laid everything out, gave it my version of events, and asked it to break down what was going on. The response came back clean, logical, and exactly what you would expect. I was right. She was wrong. It even explained why in a way that made me feel pretty good about myself.

So naturally, I brought that confidence straight to her.

And within about thirty seconds, she challenged the entire thing and gave me a completely different perspective that, if I am being honest, I hadn't really considered. So I went back, took what she said, put it into the same tool, and asked it to analyze the situation again.

Now I was wrong.

Same tool. Same situation. Completely different outcome. The only thing that changed was the input. Which, for the record, is not the outcome I was hoping to validate.

That was the moment it clicked. Claude did not give me the truth. It gave me a reflection of how I framed the problem. And if I had stopped at the first answer, I would have walked away more confident and more wrong at the same time.

Is AI making us smarter, or quietly making us worse?

That stuck with me, especially because earlier this weekend I was reading something in The Wall Street Journal by Vivienne Ming that asked a bigger version of the same question. Is AI actually making us smarter, or is it quietly making us worse?

Ming ran an experiment that was pretty straightforward. She split participants into three groups. One group used AI by itself, one group worked as humans only, and the third group combined human decision making with AI. Everyone had about an hour to make predictions about real world events using scenarios pulled from Polymarket. In theory, the hybrid group should have been the clear winner. That is the whole promise of AI, combining human judgment with machine intelligence to get better results. But that is not what happened.

The humans working alone did not perform particularly well, which is not surprising. Most of them leaned on headlines or surface level narratives and never really got past that. The AI on its own did reasonably well, but not at a level that would make you think it is replacing human thinking anytime soon. The interesting part was the hybrid group. They had the best tools available to them and still did not outperform in the way you would expect. The reason was not complicated. They were not actually using AI to think better. They were using it to validate what they already believed. They came in with a point of view, fed it into the model, and looked for support. The AI responded accordingly.

What makes this more concerning is that this is not just one isolated study. Broader research across dozens of experiments has shown something similar. In many cases, humans working with AI do not outperform the best standalone performer. Not because the technology is lacking, but because people either over trust it or underuse it. The gap is not capability, it is behavior. And unless people are given time to learn, get feedback, and actually understand how to work with these systems, they default to the easiest path, which is either blindly accepting the answer or ignoring it altogether.

The systemic risk

That is where this starts to get bigger than just people using AI wrong. This starts to look like a systemic problem if it scales.

This weekend alone, I saw a headline floating around that your future doctor is using ChatGPT to pass medical school, so maybe start eating better now. It is funny until you think about it for more than ten seconds. Because the issue is not that someone used AI to get help. The issue is what happens if they never actually understand the material behind the answer. What happens when the habit becomes skipping the reasoning and going straight to the output. At some point, that shows up in the real world, and in certain professions, that is not a small problem.

The risk is not that AI gives a wrong answer. The risk is that we stop asking the right questions. We stop asking why. We stop asking what if. We start treating outputs as conclusions instead of starting points. And when that happens, you are not solving problems anymore, you are reacting to symptoms. If you are responding to a symptom instead of the underlying issue, you might feel productive, but you are just getting further away from the truth.

There is also growing evidence that this behavior is not neutral. Early research is starting to show that heavy, passive reliance on AI can reduce critical thinking and lower engagement with problems. People analyze less, question less, and accept more. Over time, that becomes a habit. You stop thinking things through and start defaulting to whatever sounds right first.

You can usually tell when that is happening. The writing feels off, the logic does not quite hold up, and the confidence of the answer outweighs the substance behind it. Over time, that starts to hurt credibility. People may not call it out immediately, but they recognize it. And once credibility starts to slip, it is hard to get back. There is also a deeper issue underneath all of this, which is how easy it is to get AI to agree with you. If you keep pushing in a certain direction, you can usually get the model to support almost any position. That does not mean the position is correct. It just means you are guiding the response.

Curiosity beats certainty

One of the more important points in the article was around perspective taking and intellectual humility. Perspective taking is not just about hearing another side so you can argue against it. It is about actually trying to understand how someone else sees the situation and why they arrived at that conclusion. Intellectual humility is recognizing that there are limits to what you know and being willing to sit with that instead of rushing to fill the gap with a quick answer. Most people do not do either of those things, especially when they have a tool that can give them something that sounds like an answer immediately.

"All them fellas that used to belittle me, not a single one of them was curious. They thought they had everything all figured out, so they judged everything, and they judged everyone. If they were curious, they would've asked questions."

That is the entire issue with AI right now. People are not using it with curiosity. They are using it with certainty.

The people who will get the most out of AI are not the ones using it the fastest or the most often. They are the ones who treat it like a conversation instead of a shortcut. They ask better questions, follow up on the answers, and actively try to find where the response might be wrong. They use it to test their assumptions rather than reinforce them. That takes more effort, not less, but it is also where the real value shows up.

At the end of the day, AI is not automatically going to make anyone smarter. It tends to amplify whatever is already there. If someone is careless, it will make that more obvious. If someone is biased, it will reinforce that bias. But if someone is thoughtful, curious, and willing to challenge themselves, it becomes a very different kind of tool. It sharpens thinking instead of replacing it.

That is really the takeaway. In a world where you can get an answer to almost anything in seconds, the advantage does not come from having access to information. It comes from how you engage with it. The people who stay curious, who are willing to question what they are seeing, and who are comfortable admitting they might not have it all figured out yet are the ones who will actually benefit from this shift.

Oh, and one other thing. Before you run your next argument with your spouse or significant other through tools like Claude or ChatGPT and walk in feeling like you have got a bulletproof case, try putting in their point of view too. Even if you think it is completely off base. Best case, you actually gain some perspective. Worst case, you at least know what is coming and can prepare your counter. Either way, you are a step ahead of where you started.

In other words, if you are going to use AI, do not just use it to be right. Use it to understand.

Be more like Ted Lasso.