Yesterday I had a conversation that left me thinking more about how we use AI tools like ChatGPT or DeepSeek.
Someone told me about an email written by ChatGPT. It was a complaint. ChatGPT drafted it well — polished, clear, assertive. The email was sent. But the response to the complaint was dismissive. The complaint was seen as petty. Worse, the sender came off looking petty too.
That’s when it hit me: ChatGPT only does what you ask it to do but it never questions whether or not you should do it, unless you ask, “Is this really worth complaining about? Is the complaint too trivial? How does the complaint make me look?”
ChatGPT is good at putting your thoughts into words, but it never questions the wisdom of your thought, unless you specifically ask. If you don't ask, you might sound smart but look unwise.
Unless you ask it to play devil’s advocate or challenge your thinking, it will just go along with whatever direction you set. In a sense, it acts like an overly supportive friend — the kind who always agrees with you, never challenges your bias and assumptions,
What I learned is when using ChatGPT, we can’t outsource our judgment and wisdom. We have to be our own compass. Before asking it to write something, we might first ask ourselves or even ask ChatGPT:
What am I really trying to achieve?
Could this come across the wrong way?
Is there a more mature or constructive response?
ChatGPT is only as good as the questions we ask and the conversations we’re willing to have with it. Like any tool, its value is shaped and sharpened by the hand that holds it.
So before you hit send — pause. Ask better questions. Invite a challenge. And remember: ChatGPT can write your words, but it's always up to you to protect your voice.
More on this later...
Comments
Post a Comment