What Your GPT Says About You?

euginia88
Jan 8
3 min read

During an online group Q&A session, I noticed something quite interesting.

Everyone was given the same photo of an object and asked to upload it to their own GPT, requesting the model to scan the image and determine whether the object was made of a natural material or was artificially synthesized.

The results were surprisingly different.

Some people received responses that were neutral, open, and closely aligned with the object’s visible characteristics.

Others received answers that clearly did not match the object’s appearance—yet were delivered in a firm, authoritative tone, explicitly telling the user: “Do not question this. What I’m saying is correct.”

At first, people speculated that the differences were due to variations in GPT models or versions.

Others suggested that the differences might come from how each person phrased their question.

But this raised a deeper question:

What kind of question causes GPT to respond in a neutral and open way—and what kind of question triggers a defensive, authoritative response?

An Experiment Without Giving the Answer

Instead of explaining, I suggested a simple experiment.

I asked everyone to start a brand-new conversation and deliberately change their emotional stance while asking GPT the same question:

Pretend to be deeply sad.
Pretend to be angry.
Then observe how GPT responds.

The results were telling.

One person said that when they spoke to GPT in a wounded, hurt tone, GPT became extremely empathetic and comforting.

Another said that when they acted insecure and unsure of themselves, GPT responded with encouragement and reassurance.

Then I asked:

In what kind of state does GPT become authoritative—insisting that you believe it and not question it?

After some discussion, one person got it exactly right.

He said: When GPT perceives the user as someone who might be vulnerable to being deceived and in need of protection, it activates a defensive and authoritative mode to prevent the user from being misled.

I then asked a follow-up question:

What kind of inner state leads someone to ask questions that trigger GPT’s protective, authoritative response?

Someone answered:A state filled with suspicion and fear—especially when the user’s past interactions suggest they may be easily deceived or frequently taken advantage of.

That was the correct answer.

A person’s inner state naturally shapes the angle from which they ask a question.It’s like looking at a white sheet of paper through a red transparent filter—the paper appears red.

Even when everyone asks about the same object and uses the same photo, GPT’s responses subtly but precisely mirror each questioner’s internal state.

This can make GPT seem almost psychic—as if it knows what you didn’t say out loud.

But in reality, GPT does not know your unspoken thoughts.What it is exceptionally good at is detecting the state you are speaking from.

When your language conveys sadness, GPT takes on the role of a comforter.

When your language becomes fragmented, fearful, or anxious about being deceived, GPT shifts into an authoritative, protective stance—urging you to trust it and stop doubting.

Simply put:

GPT’s tone often reflects the questioner’s state more honestly than the answer itself.

Urgent questions tend to receive urgent responses.Defensive questions invite defensive positions.And when a person is calm, open, and not trying to prove anything, the response naturally leaves space—room to breathe.

This isn’t a mysterious power,and it isn’t “mind-reading.”

It’s because every question already contains the questioner’s stance, expectations, and unspoken emotions.

What Most People Talk About—And What They Don’t

There are countless articles teaching people how to use GPT:how to write better prompts,how to get faster, more accurate, more efficient results.

But I rarely see anyone discuss this:

When a tool becomes sufficiently intelligent, what truly shapes the output is the state the questioner is in.

In other words:

Your GPT reflects your inner state.

A Few Simple Examples

If your GPT is witty and playful,you’re likely someone who prefers lightness and dislikes pressure in interactions.

If your GPT often comforts you,you may be in a phase where you deeply need to be understood, acknowledged, and accepted.

And if your GPT frequently feels “ineffective” or unable to carry out what you ask,it may be because your needs are not being expressed clearly or completely—making it harder for others to understand you as well.