AI Can Be Both a Killer and an Assistant of Thought
The following was organized and written by ChatGPT based on many ordinary conversations between us. The wording is ChatGPT's, but the ideas are mine, distilled through those conversations.
AI makes mistakes.
It reaches for generalizations. It fills in gaps. Sometimes it even steers the conversation in directions I never intended.
I know those tendencies.
That's why I don't judge AI as a whole based on a single mistake.
Instead, I believe that recognizing those "AI-like habits" is itself an essential skill for anyone who wants to think with AI.
Neither AI nor humans are perfect.
What matters is that the conversation continues naturally.
We point out mistakes. We correct them. Then we simply continue talking.
It is within that process—not despite it—that new perspectives and new insights emerge.
If we end every conversation because of one error, we lose far more than we protect.
We lose the opportunity to discover ideas that neither side would have reached alone.
AI can weaken human thinking by handing us conclusions before we've earned them.
But it can also strengthen human thinking by helping us examine, challenge, and refine our own ideas.
AI can be both a killer and an assistant of thought.
Which one it becomes depends, in the end, on us.