I’m not sure whether I’m more impressed by the technical proficiency of the software developers behind ChatGPT — or if I’m more alarmed by the dystopian future that I see such software leading to.
What we call “artificial intelligence” is nothing but software. It isn’t intelligent. It has no consciousness. It has no actual awareness or understanding of what it produces. It’s just lines of computer code written to produce material that mimics human behavior. If you think of AI as some form of semi-consciousness, you’re buying into science fiction. This is nothing but software written by clever people — and it’s nowhere near as “smart” as you’ve been led to believe.
But AI software — such as ChatGPT and its competitors — is getting better and better at spitting out content that mimics what a human might have created with real thought. And I think this is dangerous.
As an experiment, I asked ChatGPT to create an essay in my own writing style. I didn’t give it a subject. This is the only instructions I gave the software: “Write an 800-word essay in the same style used by the writer of davidmcelroy.org.”
The results shocked me.

‘Cash for clunkers’ was an even bigger clunker than we first realized
We have no choice but to trust even in face of betrayal and hurt
Hospital’s five-year fight to move shows health care isn’t free market
Donald Trump is an evil man, but his political enemies are evil, too
Trusting Obama to create jobs is like trusting an arsonist to put out fires
Apple podcast listing means you can now subscribe to Love & Hope
It’s hard to live in tension between what we ‘know’ and the unknown
Maybe it’s easier to do hard things when nobody says they’re difficult