AI is amazing. But if it makes you stop thinking, that is not leverage. That is just outsourcing your judgment.
I am seeing this more and more at work. People asking AI for answers and taking those answers as the truth. Word for word. No follow-up. No second pass. No comparison against their own read of the situation. They get an output, and they ship it. That is not how this is supposed to work.
AI is supposed to help you analyze. It is supposed to make you faster. It is supposed to give you a framework. But the actual answer, the call, the judgment, that is still yours. You are still the one accountable for whether it is right. The tool does not absorb the consequences. You do.
Use AI as a tool. Do not become the tool.
The quiet erosion of thinking
That distinction sounds small until you look at what is happening when people forget it.
What is happening is a quiet erosion of the actual thinking work that used to define professional competence. The reasoning. The pattern matching. The willingness to sit with ambiguity for an extra ten minutes before reaching for an answer. All of those muscles get smaller every time someone hands the problem to AI without engaging with it first. The output looks identical. The reasoning behind the output is gone.
The data is not flattering
There is now hard data on this, and the data is not flattering.
A 2025 peer-reviewed study published in the journal Societies surveyed 666 participants across diverse age groups and educational backgrounds in the United Kingdom on AI tool usage, cognitive offloading, and critical thinking. The findings were direct. There was a strong negative correlation between frequent AI tool usage and critical thinking ability. Specifically, AI tool use correlated with critical thinking at minus 0.68. Cognitive offloading correlated with critical thinking at minus 0.75. Translation. The more people leaned on AI to do their thinking, the worse their thinking got. And the more they offloaded the cognitive work itself, the steeper the decline.
That is not a marginal effect. Those are some of the strongest negative correlations you will see in social science research. The relationship is real. It is measurable. And it is happening right now in offices, classrooms, and decision-making rooms across every industry.
The same study found that younger participants, the 17 to 25 year old cohort, showed the highest AI dependence and the lowest critical thinking scores. Older participants, the 46 and over group, showed the inverse. Lower AI dependence, higher critical thinking scores. Higher education levels mitigated some of the effect, but did not eliminate it. The pattern is the pattern.
I am not saying AI is bad
I want to be careful here. I am not saying AI is bad. I am not saying do not use AI. I use AI every single day, and the work I do with AI is sharper, faster, and more leveraged than anything I could produce without it. The argument is not against AI. The argument is against a specific way of using AI that is becoming common and is doing real damage.
The wrong way to use AI is to hand it the question and accept the answer.
The right way is to hand it the question, read the answer, push on the answer, ask why it gave you that answer, ask what is missing, ask what it might have gotten wrong, and then arrive at your own answer informed by all of it. That is leverage. That is what these tools are designed for. They are designed to make a thinking person more powerful, not to replace the thinking entirely.
What happens when you skip the thinking
When you skip the thinking, two things happen at the same time. The first thing is that you accept whatever bias, hallucination, or shallow read the AI happens to produce. AI is not always right. Anyone who has used these tools seriously knows that they confidently say wrong things on a regular basis. If you do not check, you ship the wrong answer with the same conviction the AI delivered it. The second thing that happens is more insidious. Your own judgment muscles atrophy. You lose the ability to spot when something is off, because you stopped practicing the spotting. You lose the calibration that comes from making your own calls and being right or wrong about them in public. Every shortcut compounds. Eventually you do not have the judgment to evaluate what AI gives you, even if you wanted to.
I see this happening with younger professionals especially, and the research backs that up. The 2025 cohort study is the cleanest signal we have. Younger workers raised on AI tools have a different relationship with their own thinking than older workers do. That is not a moral failing. It is a function of having had AI available for the entirety of their professional life. The risk is that the convenience hardens into dependence, and the dependence becomes the operating mode. By the time anyone notices, the muscle is already gone.
The fix
The fix is simple in concept and hard in practice.
Treat AI like a research assistant, not an oracle. A research assistant gives you a draft. You read the draft. You catch the errors. You add the context the assistant did not have. You make the call yourself. The assistant did real work, but the output is yours, and the accountability is yours, and the thinking is yours.
An oracle gives you the answer, and you accept it because you trust the source. That posture works fine for trivial things. What time is it. What is the weather. How long is the drive. It is a disaster for anything that actually requires judgment, because the second you stop checking, you stop being the one who is thinking.
Who wins this decade
The professionals who win this decade are not going to be the ones who use AI the most. They are going to be the ones who use AI most thoughtfully. The ones who keep their own judgment sharp by exercising it on every output AI gives them. The ones who treat the tool as an amplifier of their thinking, not a substitute for it.
The real estate parallel
In real estate, I see this distinction every week. There are agents using AI to draft listing descriptions, summarize market data, and surface client patterns. The good ones read everything AI produces and revise it. They catch the language that does not match their voice. They notice the data point that is off. They remove the framing that does not fit the local market. The output is better than what they would have produced alone, and the thinking is still entirely theirs. Those are the agents who are pulling ahead.
Then there are agents who hand the work to AI and ship the output unchanged. That work is identifiable from a mile away. Generic. Off-tone. Sometimes flat-out wrong. Their clients can feel it. Their colleagues can feel it. The market can feel it. The shortcut is visible to everyone except the agent who took it.
The agents shipping unedited AI output also tend to be the ones who cannot tell you why they made the choices in the document. Ask them why a particular phrase was used, and they cannot answer. Ask them why one comparable was emphasized over another, and they shrug. They stopped being the author of their own work. They became a delivery mechanism for whatever AI produced. That is not leverage. That is replacement, dressed up as productivity.
The agents who are revising every output are doing something different. They are using AI to handle the first draft, the rough cut, the structural scaffolding. Then they are doing the actual professional work, which is the judgment about what fits this client, this property, this market. Their AI use does not reduce their accountability. It concentrates it. They are responsible for everything that goes out, and they know it, and they engage with it.
The bottom line
This applies to every industry. The pattern is the same. The leverage goes to the people who are still thinking. The cost falls on the people who stopped.
AI is a tool. The most powerful tool I have ever had access to as a professional. The same way a calculator is a tool, the same way a spreadsheet is a tool, the same way the internet is a tool. None of those replaced the thinking. They amplified it. AI is no different. Used correctly, it makes serious thinkers more serious. Used incorrectly, it lets shallow thinkers stay shallow without realizing it.
The question every professional should be asking themselves right now is which of those two patterns describes their relationship with AI. There is no neutral middle ground. You are either using the tool to amplify your thinking, or you are letting the tool replace your thinking. Both feel productive in the moment. Only one of them is.
Use AI as a tool. Do not become the tool.
Judd Hoffman is CEO and Co-Founder of Ethica AI, building AI-powered tools for real estate transaction workflows.