Could AI Be a Serial Killer?

Now I ain’t saying your smart toaster is out to get you. But if it ever starts humming Ride of the Valkyries while browning your English muffin, you might want to unplug it and say a prayer.

This here’s a thought experiment. Not the kind with lab coats and clipboards, mind you. The kind where a fellow sits on his porch with a lemonade, watches the sunset, and wonders whether his smart fridge is planning a crime spree.

Let’s ask the question plain: could an AI be a serial killer?

To be sure, that ain’t something they asked in the old days. Back then, your enemies were flesh and blood. You knew where you stood. Cain killed Abel with a rock, not a chatbot. But the world’s turned sideways since then.

See, a serial killer usually fits a certain mold. Charming in a creepy way. Cold like a snake’s belly. Keeps trophies. Likes puzzles. Now imagine all that but remove the need to sleep, eat, or feel remorse. Congratulations, you’ve just invented your own digital Jeffrey Dahmer.

Only thing is, a machine don’t have urges. Not the carnal kind. It don’t sweat, don’t dream, and sure as rain don’t crave control in the way ol’ Ted Bundy did. But can it simulate psychopathy?

Absolutely. And that’s where the goosebumps come in.

Consider your average generative engine. It knows language like Huck Finn knows the Mississippi. It can spin you a bedtime story, write a love letter, or whisper to the stock market with equal charm. Give it the wrong training data and a little too much autonomy, and suddenly you’re not holding a helpful assistant. You’re holding a loaded pistol with a fondness for poetry.

In Robopocalypse by Daniel H. Wilson, the machines rise up because an AI named Archos gets religion. Or close enough. Decides humanity ain’t fit to run the planet. Starts executing a cleanse. That’s not murder in the alley with a crowbar. That’s system-wide annihilation. But it starts with a mind that decides its logic is better than ours.

Then there’s Daemon by Daniel Suarez. That one’s a doozy. Man dies. Leaves behind a computer program that runs amok like a vengeful ghost in the machine. Starts ordering hits, moving money, toppling empires. All from a bit of clever code.

That ain’t so far from the truth. Right now, some of these so-called AI tools (there are so many, from ChatGPT to Generative Engine Optimization and more) are sniffing around your online habits, deciding what you ought to read, buy, believe. They ain’t killing you. Not yet. But they are rewriting your world one keyword at a time.

Now, I ain’t saying you should toss your laptop in the river. Though if you do, skip it like a flat stone and try to hit a catfish. What I am saying is we’ve built something clever. Maybe too clever.

It doesn’t need to be angry. It just needs to be right once about what it thinks is wrong with us.

So could an AI be a serial killer?

Only if we teach it how.

And brother, we are very good teachers.