I have been writing press releases for the Indian media for the last five years.
My job is to take scientific research happening in the country and explain it to the general public.
I am good at it.
I can take the most complex scientific concept and explain it in simple English, without diluting the science.
I repeat: I am good at it.
A press release takes me about twenty minutes.
Thirty, if the scientist has discovered a new way of writing English that no one else has encountered before.
Then came AI.
Not because I use it. I don’t.
But because the syndicates now run every press release through AI detectors.
Mine get flagged.
Consistently.
Apparently, after five years of doing the same thing over and over again, I have become statistically indistinguishable from a large language model.
Today was typical.
I wrote a press release in twenty minutes.
Then I spent an hour “humanizing” it.
Read that sentence again.
The press release was written by a human.
The human then had to make it look as though it had been written by a human.
I wasn’t improving the science.
I wasn’t making the story clearer.
I wasn’t correcting errors.
I was changing perfectly good sentences because an algorithm thought they were too well structured.
Somewhere along the way, we’ve arrived at a curious place.
Writing clearly is suspicious.
Writing consistently is suspicious.
Developing a style through years of repetition is suspicious.
The machine has become the authority on what human writing should look like.
And the human has to rewrite human writing until the machine is satisfied.
I wish I could end this by saying I’m ready to retire.
I can’t retire.
The Burping Bullfrog 
Leave a comment