There is a particular kind of conversation I keep having, and I have started to dread it.
Someone finds out I write about testing AI tools, and within two minutes they ask me a question that has nothing to do with tools. They ask me whether I think AI is good. Whether I trust it. Whether I am "for" it or "against" it. Sometimes they ask whether I think it is dangerous, or whether I think it is overhyped, or whether I think it will change everything.
I understand why people ask. But the question assumes something I do not accept: that AI is a thing you have a relationship with. A stance. An identity.
It is not. It is a tool. And treating it like a personality test is the fastest way to stop thinking clearly about it.
The Moment This Became Clear to Me
A few months ago I was in a meeting — the kind where someone shares a screen and the conversation drifts — and a colleague asked the room, "So are we pro-AI or anti-AI as a team?"
Nobody answered. The silence was not because people did not have opinions. It was because the question was the wrong shape. It asked for a position where what was needed was a decision about a task.
I have thought about that silence a lot since. It is the exact moment my approach to this whole subject snapped into focus. Nobody asks whether a team is "pro-spreadsheet." Nobody has a personal relationship with a stapler. You use the tool when it helps and you put it down when it does not, and you do not feel anything about it either way.
That is the stance I try to write from now. Not pro, not anti. Just specific.
Three Ways the Personality Test Framing Goes Wrong
It turns a task question into an identity question
When someone asks whether I "trust AI," they are not asking about a specific tool doing a specific job at a specific time. They are asking me to declare a position. But trust is not a property of AI. It is a property of a relationship between a task, a tool, and a person.
I do not trust AI to write my final copy. I do trust it to generate twelve headline options that I will then throw away. Same tool. Different task. Different answer. The personality-test framing cannot hold that nuance, so it collapses it into "do you trust it or not," which is not a useful question.
It makes disagreement feel like conflict
Once you frame AI as something people are "for" or "against," every conversation becomes a debate about allegiance. Someone who had a bad experience becomes "anti-AI." Someone who had a good one becomes "a believer." Neither label describes what actually happened, and both make it harder to say the truthful thing, which is usually "it depends on the task."
I have watched people soften honest criticism because they did not want to be seen as taking a side. I have watched people oversell a tool because they did not want to be seen as behind. Both are the personality test doing damage.
It hides the only question that matters
The useful question is never "is AI good." It is: for this task, on this day, with this input, did the tool help, and what did it cost me to find out?
That question has a specific answer every time. Sometimes the answer is yes and it saved me twenty minutes. Sometimes the answer is no and it cost me an hour I will not get back. Sometimes the answer is "it helped with the first draft and I rewrote most of it." All of those are real, and none of them are a personality.
What I Actually Do Instead
I have a short set of questions I run through before I use a tool for anything, and they have nothing to do with how I feel about AI.
What is the task? Not the category — the actual finish line. "Draft a weekly report" is a category. "Turn these eight bullet points into a first draft I can edit" is a task.
What does a good result look like? If I cannot describe it, I cannot judge whether the tool helped.
What will I have to check? Every tool has a checking cost. If the checking costs more than the drafting, the tool is not helping.
What happens if it is wrong? Low stakes, high stakes, or something in between. This changes everything.
Would I use it again for this exact task? Keep, use selectively, wait, or cancel.
Notice that none of these questions require me to have an opinion about AI as a concept. They only require me to pay attention to the task in front of me.
That is the whole method. It is not exciting. It is the thing that actually works.
The "For or Against" Trap on Both Sides
I want to be fair here, because the personality test is not only a problem for skeptics.
The hype side does it too. "AI will change everything" is a stance, not an observation. It asks you to position yourself in a story about the future rather than to look at what a tool did on your desk last Tuesday. It is just as unhelpful as "AI is overhyped," and it is just as easy to get swept up in.
I have been guilty of this in smaller ways. Early on, when a tool worked well, I wanted it to mean something bigger than it did. A good session became evidence that I was "on the right side of this." That is not testing. That is rooting.
The correction is the same on both sides: get specific. Name the task. Name the result. Name the limitation. Let the tool be a tool.
What This Means for How I Write Here
Every article on this site is an attempt to answer a task question, not a stance question. That is deliberate.
When I write "I Tested Five AI Writing Tools and Kept One," I am not saying AI writing tools are good or bad. I am saying four of them did not earn their place in my workflow for the tasks I gave them, and one did. That is a test result, not a position.
When I write "I Did Not Save Three Hours. I Saved the Worst Twenty Minutes," I am not saying AI is disappointing. I am saying the tool did something smaller and more specific than the marketing promised, and that smaller thing was still worth having.
The signature phrase on this site — "I tried it so you don't have to waste your afternoon" — is deliberately not a stance. It is an offer. It says: here is what happened when I ran the test. You can decide what to do with that.
The Limitation to Remember
AI is not a personality test. It is not a team you join. It is not a side you pick.
It is a tool that does some tasks well, some tasks badly, and some tasks in a way that looks impressive until you check the result. Your job is not to decide how you feel about it. Your job is to decide, task by task, whether it earned its place in your workflow — and to be honest when it did not.
The moment you turn it into an identity, you stop being able to answer that question clearly. And clear answers are the only thing this site is actually trying to provide.
The Anonymized Version of This Practice
If you find yourself in a conversation about whether you are "for" or "against" AI, try redirecting it with a task question:
"I do not really have a position on AI in general. But I can tell you what happened when I used it for [specific task]. Want to hear that?"
It is a smaller conversation. It is also the only one that produces anything useful.
Guardrails I Followed
I did not paste any employer, client, or coworker details into any tool to write this piece. The meeting described is anonymized and generalized.
I did not present my task-based approach as the only valid one. It is the one that works for me, and I said so.
I did not use fear, urgency, or a claim about the future to argue for or against AI adoption.
I kept the focus on tasks and results, not on identity or allegiance.
The stance question came from somewhere else. The task question is mine.
Next in this series: I used AI to sort a month of content ideas. Here is what it missed.
I tried it so you don't have to waste your afternoon.
No notes yet — be the first to inscribe one.