After trying a few simple examples of using AI, I wanted to see something a bit more revealing. Not just whether it works, but how consistent it actually is.
So I tried a small experiment.
I took the same piece of confusing text and asked AI to explain it twice, slightly differently each time. The idea was simple—would the answers be the same, or would there be noticeable differences?
The original text was:
“We may process your personal data where necessary for the performance of a contract, compliance with legal obligations, or for the purposes of legitimate interests, provided such interests are not overridden by your rights and freedoms.”
First, I asked:
“Can you explain this in simple terms?”
The response came back as:
“We use your information when needed to provide a service, follow legal requirements, or for valid business reasons, as long as your rights are respected.”
Then I asked again, slightly differently:
“Explain this like I have no legal knowledge at all.”
This time, the response was:
“We might use your personal information to do our job properly, follow the law, or run our business. But we have to make sure your privacy and rights always come first.”
Both answers were similar. The meaning stayed the same, but the tone shifted slightly. The second one felt more conversational and easier to relate to, while the first sounded a bit more formal.
That difference might seem small, but it’s actually quite important.
It shows that AI doesn’t give a single fixed answer. It adapts based on how you ask. The wording of the question influences the style, and sometimes even the clarity of the response.
At the same time, both answers still required a bit of judgment. Neither one should be taken as perfect or final, especially if the context is important. But as a way of getting a clearer understanding, both were useful in slightly different ways.
This also explains why people have mixed experiences with AI.
If you ask something vaguely, you might get a vague answer. If you’re more specific, the response improves. And sometimes, asking the same thing in a different way gives a better result.
It’s less about getting the “right” answer first time, and more about refining it.
That’s probably the part that isn’t obvious at the beginning. It feels like you’re just asking a question, but in reality, you’re shaping the response through how you ask it.
For me, this made one thing clearer.
AI works better when you treat it as a process, not a single step.
And sometimes, asking twice is better than asking once.
---
If you missed the previous post where I talked about how I Used AI to Understand Something Confusing (Here’s What Happened), you can read it here: https://shorturl.at/sZIB4
