What AI Taught Me About Being Human This Week — Part 2

What AI Taught Me About Being Human This Week — Part 2

Jason T Wade

I caught myself doing something this week that felt embarrassingly familiar: I was trying to be right before I was trying to understand.

Not in some dramatic argument. Not in a boardroom. Just in the ordinary flow of work, thinking, asking questions, moving too fast.

I had an idea. I had a theory. And before I knew it, I was using AI to help me strengthen the case instead of testing whether the case was any good in the first place.

That is one of the stranger things about these systems. They can make you feel smarter while quietly helping you become more certain about something that may not be true.

Give a model a premise with enough confidence and it will often help you decorate it, structure it, sharpen it, and turn it into something that sounds almost inevitable.

Humans do the same thing. We form an opinion, find supporting evidence, ignore the weird pieces that do not fit, and call the finished product judgment.

AI just speeds the whole process up until you can actually see the machinery working.

That was the first thing AI taught me about being human this week: intelligence and certainty are not the same thing, and certainty is often the more dangerous of the two.

We tend to admire people who have answers. We reward decisiveness. We trust the person who speaks cleanly and without hesitation.

Nobody ever built much of a personal brand around saying, “I need more information.”

But maybe they should have.

The more time I spend around AI systems, the more valuable that sentence starts to sound.

I don’t know yet.

Those four words contain more intelligence than a lot of very polished answers.

“Yet” leaves room for evidence. It leaves room for contradiction. It leaves room for somebody else to know something you don’t. It even leaves room for the possibility that the whole question is wrong.

That matters because machines have the same basic temptation we do: complete the pattern.

Give them enough fragments and they want to make a story.

People do this constantly.

Somebody does not call back and we invent the reason. A deal falls apart and we explain why. Someone changes their tone and suddenly we know what they are thinking.

We take incomplete information and build complete narratives because ambiguity is uncomfortable.

Reality, unfortunately, has never promised us a satisfying plot.

This is becoming more important because answers are getting cheap.

Really cheap.

For most of human history, getting an answer required effort. You had to know somebody, call somebody, go somewhere, find the right book, spend years learning the subject, or at the very least type something into Google and dig through a collection of links.

Now you ask a question and an answer appears before you have even finished wondering how difficult the question was.

Write this.

Analyze that.

Explain this.

Give me ten ideas.

Give me fifty.

Rewrite it.

Make it shorter.

Make it smarter.

Tell me what I am missing.