I asked AI bot ChatGPT to write a column on happiness. Here’s what happened next

I’ve been dying to know. Am I going to be replaced? Does the arrival of an artificial intelligence chatbot such as OpenAI’s ChatGPT mean my days as a columnist are numbered?

I had to find out for myself.

So I conducted a little experiment. I asked ChatGPT to write an article for me.

I submitted the prompt, “Write a 500-word column on happiness.” That’s a broad, overworked topic, I know. Still, I wanted to see what it could do.

My impression of the final product? A resounding “meh.”

The piece definitely had its strong suites. Topping the list was its mind-numbing speed.

In the length of time it usually takes me to type a title and byline, ChatGPT had pumped out an entire column. Wow.

The content was okay. It covered the major bullet points about the assigned topic.

It presented them in a well-organized manner, using solidly- constructed paragraphs and grammatically correct sentences. The punctuation was spot on.

A high school English teacher would be pleased.

That’s where my admiration waned. I then felt the urge to don my red pen and edit like crazy. The writing was tepid at best — accurate and logical but boring.

Let’s start with the repetition. The 500-word article used the terms “we tend” and “relationships” seven times each, and “the truth is” and “the key is” three times.

It obviously hadn’t consulted a thesaurus or considered changing things up.

Next, the ChatGPT article needed punch. The sentences were too long. The language was mundane.

The column read like it came from a robot. Nothing in the article made me smile, grapple with a statement or feel enraged. It was soul-less, devoid of personal connection.

If the column had been a pair of shoes, Marie Kondo would have advised me to ditch them because they didn’t spark any joy.

Of course, what was I expecting from something created by artificial intelligence?

If the piece wasn’t going to be personal, at least it could have been scientific. However, no research was cited, and no experts chimed in.

Nothing encouraged me to trust the information other than the fact that it was printed on the page.

To add insult to injury, the article began the final paragraph with the words “in conclusion,” the most yawn-worthy way to sign off.

I always try to end big — leaving readers wanting more or at least appreciating something they’ve read.

The robot obviously doesn’t share my sentiments, choosing to close by recapping its humdrum findings instead.

Don’t get me wrong. I don’t dislike ChatGPT. I’m gobsmacked by its speed.

I see it being useful for producing a quick rough draft that gets fine-tuned to the writer’s whims. I’m sure I could refine my search and get a piece that’s more to my liking, by asking the bot to cite three scientific studies or use bouncier language.

This was my first cautious meeting with AI, to see if it posed a threat. And while I think this newcomer has massive potential, my job appears safe for now.