Humans + AI

Humans + AI think, communicate, decide, learn and adapt better together.

I’ve been thinking about how AI is changing what I expect from people, not because I expect people to behave like AI, but because it changes what should count as differentiation.

I’m a systems thinker. It’s how I solve problems and, increasingly, how I communicate: What are we trying to accomplish? What are the constraints? What assumptions are we making? What affects what? What happens downstream?

Before AI, simply being able to structure a complicated problem and reason through it well could be a significant differentiator.

But increasingly, AI can help almost anyone do that.

So I’ve found myself thinking:

If access to better thinking is becoming abundant, what should I expect from people?

My first instinct was: perhaps better thinking itself is no longer the differentiator. Instead, I should expect people to communicate clearly and actually use structured ways of thinking when solving problems.

I explored that thought with ChatGPT, and it challenged the premise in a useful way.

It suggested separating three things:

Thinking → articulation → judgment.

AI is rapidly raising the baseline for the first two.

But judgment is different.

Judgment means deciding which problem actually matters. Recognizing when a perfectly coherent model doesn’t fit reality. Understanding people and organizational dynamics. Choosing between legitimate competing objectives. And ultimately taking responsibility for the decision.

That led me to another question:

If AI increasingly does the thinking and articulation, how do we develop good judgment?

The answer that resonated with me was that judgment develops through calibration:

Situation → model → prediction → decision → consequence → learning.

And that’s where our relationship with AI becomes important.

If the loop becomes:

Problem → AI thinks → AI recommends → I approve

I’m not necessarily developing judgment. I may simply be getting better at recognizing answers that sound reasonable.

A more useful loop might be:

Problem → I form a view → AI challenges it → I decide → reality responds → I compare the outcome with my prediction → I update my model.

That changes how I think about using AI.

Maybe the scarce skill isn’t simply thinking faster or having better answers.

Maybe it’s knowing what thinking to delegate, what thinking to retain, and having enough contact with reality to develop judgment.

And perhaps the differentiator for a systems thinker is changing too.

It may no longer be enough to be the person who can see the system.

The higher-order skill may be the ability to create a system in which humans + AI think, communicate, decide, learn and adapt better together.

That’s the capability I’m becoming much more interested in.

AI disclosure: The initial observations and questions in this post are mine. I used ChatGPT (OpenAI) as a thinking partner to challenge the premise, develop distinctions, and articulate parts of the argument. I reviewed and shaped the resulting piece and take responsibility for the ideas I chose to present.