In the recent past, I have been observing something that does not actually make sense. The instruments that purport to identify any piece of writing as having been done by a human being or an AI are going out of control. And today, I conducted a little experiment, which made me understand the unreliability of such tools.
My content is normally written out. Every thought, every word, every illustration– absolutely my own. Thereafter, I just correct minor grammar errors or punctuation marks using ChatGPT. I even say it to the point: “Do not modify my tone. Don’t rewrite the style. Just clean the errors.”
And so I did the same routine to-day. I created my work, requested GPT to fix small errors and then, simply out of curiosity, I ran it through an AI-based detector.
Guess what?
It showed 100% AI-generated.
And that’s when I thought, What???
Since the whole content was my own. GPT barely touched anything.
So I tried another test.
This is the first thing that I wrote entirely without the assistance of AI as well as without any proofreading. Just raw writing. After which I ran it again through the same detector.
This time it indicated 58% AI-generated.
Now I understood something significant:
It is not AI that these detectors are detecting. They are detecting style.
If your writing is:
too clean
too structured
too grammatically correct
too fluent
…then it is automatically written by AI as per these tools.
And if your writing has:
small mistakes
natural flow
uneven rhythm
a few imperfections
then it turns miraculously into human.
So now the question is:
What are the ways through which human writing is judged by recruiters, teachers or clients?
When natural writing can be identified as AI, and AI-edited writing is scored 100 percent generated, then what is the standard?
The truth is simple:
It is impossible to escape AI detection 100%.
Even if you write “How are you?” or even deliberately spell words wrong such as How to doing, there will be some thing that the detectors will pick.
This will never give a zero margin of detection, as these tools are not tools of measuring intelligence, they are tools of measuring patterns.
And in the near future, the world will not have any other option but to accept the fact that AI-assisted writing is normal. It is simply like Grammatically, spell-check, or auto-correction. We cannot say that these tools do not exist anymore.
So what can writers do?
My advice is simple:
Don’t doubt your writing.
Not to write flags, do not attempt to write less clean.
There is no need to worry that it can be detected 100 percent, it is impossible to evade.
Find your voice, concentrate on value, and originality. That’s what matters.
The detectors do not define your authenticity, as long as your ideas are yours, your thoughts are yours, and your message is yours.
Since it is not the grammar that is the real human touch.
It is in the purpose, the view and the experience of what you say.
And none of the detectors in the world can gauge that.
