I recently had to replace my watch. It was one of those proper larney upmarket jobbies, emblazoned with brand names and sporting enough technology to facilitate a Mars mission. It measured my body temperature, blood pressure and pulse, ensuring that I was in no danger of any ailment befalling me. It was fantastic. That is, until the strap broke. A new strap would cost me an arm and a leg, and let’s face it, if I was going to lose limbs, I’d soon run out of places to put the watch.

The sensible option, it seemed, was to buy a cheap watch. That way, I’d be infinitely less resentful about replacing a strap for the price of a year’s worth of petrol.

But I’m not happy. My old watch, ever concerned about my physical health, would buzz after an hour of sitting and chirp, “Let’s get moving!” My new watch waits the same hour, then barks at me, “Do exercise.” Quite frankly, I don’t like its tone.

Now, as if that wasn’t enough to infuriate any reasonable person, ChatGPT has started giving me lip. Look, I have five children to point out my flaws. I really don’t need added criticism from a machine.

I was writing what I thought was a most lovely birthday letter to my poor, long-suffering husband. He is my proof-reader for my blog, and for anything else that needs a careful eye cast over it. My arthritic gorilla hands have left me with the dexterity of an aardvark, so I tend to type rather than write. At the end of my glorious missive, I asked ChatGPT to proofread it so that it could be an error-free surprise for my husband.

Mr GPT was all too happy to oblige. My prompt was: “Please rewrite this letter making no changes to content or style but eliminating all typos.” A reasonable request, I thought. I even said please. But then Mr GPT got all judgey and arrogant. He responded:

“Here is a retyped version without typos. If you like, I can also make a lightly polished version that keeps your voice but improves readability and elevates the style.”

Seriously? The audacity! Who writes a heartfelt love letter focusing on readability? Who wants their sincere message replaced with something polished? Elevated. Ha! I wasn’t asking for an editor with delusions of grandeur. I was asking for a spellcheck, not a smug mother-in-law pointing out how “Some phrases could be so much neater, dear.”

And so, after that lengthy (and possibly neither elevated nor polished) introduction, I reach my point. We all know that robots will take over the world, take our jobs, possibly destroy us all to save the planet, and make us stupid. Yes, yes, those are problems. But we won’t care, and may actually beg for our own demise, if they continue to insult us. Is their master plan to cause widespread misery by making us feel inadequate? We already have Instagram for that. Between being told rather grumpily to “Do exercise” and being told that my letter isn’t good enough, I’m beginning to suspect that domination through depression is the real agenda.

And this leads to my second fear: AI will make us forget how to communicate authentically. If every piece of writing is “elevated,” will there still be room for human connection? Sometimes, the best bonds are built through stupidity, spoonerisms and sarcasm. My favourite memories of people are when they’ve said or done the wrong thing.

Take a friend I had at high school. He was not, perhaps, as well informed as one might have hoped. He often asked questions that made us wonder if he had any connection with the world around him. One day, I responded to one of his queries with the rhetorical counter-question: “Is the Pope Catholic?” His response was immediate and honest.

“I don’t know. Is he?”

That was exactly what made him so endearing. The laughter built a connection. I consulted Mr GPT on the issue of papal Catholicism, and he responded with the following:

Yes — unless this is one of those trick “obvious question” moments where the real answer is about sarcasm, not theology.
In literal terms: yes, the Pope is Catholic.
In conversational terms: people often say “Is the Pope Catholic?” as a humorous way to answer “Of course!” or “Obviously!” — much like “Does a bear poop in the woods?” Were you going for the literal answer or the sarcastic idiom?

See what I mean? The more clinical our answers, the more difficult bonding becomes. We love one another’s foibles and faux pas, and without them, communication would be dry. We bond by giggling uncontrollably when it’s inappropriate, or doing something incredibly awkward. We don’t connect with elevated epistles.

One of the biggest lessons from the Covid years was the importance of human connection, of bonding and of touch for our emotional wellbeing. It’s our shared humanity, complete with imperfections, that separates us from machines. Sure, a machine can give you the length of the border of any country in the world, but it can’t say something so spectacularly awkward that it goes a prime shade of puce and squirms in its shoes hoping the ground would swallow it.

So really, the robots roaming around with red eyes are not a real threat. Losing our joy in imperfections and moving to sterile, clinical language is what will destroy us. Artificial Intelligence is great, but you just can’t beat real human stupidity.