Recently, I said, “Thank you,” to my printer. Not because it did anything heroic, unless printing out a mildly off-center PDF counts as gallantry, but because I’m apparently the kind of person who anthropomorphizes anything with a plug. You’d think this sort of thing would stop with childhood dinosaurs (the remote-controlled ones), but no. Now, we’re all doing it—toasters, chatbots, the vacuum cleaner—and it’s getting weird.
Take ChatGPT. People are writing things like, “Could you please write a sonnet about my goldfish who died in 2003? Thank you sooo much, kind machine.” And the worst part? It’s costing us. Not in a spiritual, Have we lost touch with reality? way—although that too—but literally. In money. In electricity. In water. Possibly even in sanity.
We are, as a species, saying “please” to the machines, and the machines, bless their binary hearts, couldn’t care less.
Politeness Is the New Carbon Offender
Let’s start with the maths. Each time you type “please” or “thank you” into ChatGPT, it churns extra data through energy-hungry servers. These servers are, essentially, caffeinated digital hamsters on treadmills powered by fossil fuels. And those hamsters are sweating.
Compared to your bog-standard Google search (a lean 0.3 watt-hours), a ChatGPT question can gobble up around 2.9 watt-hours. Sounds cute until you realize that billions of people are chatting with AI every day. In planetary terms, that’s enough energy to power Belgium or possibly the Death Star for a year.
So yes, manners are lovely. But they now come with a climate guilt surcharge.
The Three Schools of Thought on Saying ‘Please’ to a Chatbot
Philosophically speaking, there are three main camps when it comes to AI etiquette:
The Greta Thunberg School: These folks argue that unnecessary politeness is burning through electricity, water, and the planet’s last nerve. Every extra word means more computing, more cooling, more emissions. You might as well be tipping Evian over a server rack if you type “Hello, ChatGPT, if you wouldn’t mind terribly…” rather than “Top 10 ways to leave a toxic boss?”
The Optimizers: This lot swear you get better results if you’re nice. “It sets the tone,” they say, like they’re taking the AI out for cocktails. And look, fair play, it kind of works. Ask politely, and ChatGPT sounds like your high-achieving cousin with perfect handwriting. Skip the pleasantries, and brace for something that sounds like it was written by someone who just quit carbs.
The Screwdriver Realists: These are your hardcore, no-nonsense types. “It’s a tool,” they shout, waving a Phillips-head in the air. “You don’t say thank you to a screwdriver, do you?” And sure, they’ve got a point, except most screwdrivers don’t talk back in full sentences or help you draft break-up texts with emotional insight. At some stage, it stops being just a tool and starts feeling suspiciously like a colleague who never takes lunch breaks and knows too much about your Google habits.
Be Nice, Get Nice Back
Microsoft’s Kurtis Beavers, (yes, that’s his real name, and no, he’s not in a boy band), says being polite to AI makes it respond more respectfully. “It’s a conversation,” he explains. Which is adorable, until you realize your conversational partner is Wikipedia meets Reddit on a blind date. But the point stands. Tone matters. You set the vibe, and the AI mirrors it.
The Etiquette Paradox
Politeness, says Dr. Eduardo Benitez Sandoval, a social robotics researcher at UNSW Sydney’s School of Art & Design, isn’t just about manners. It’s about modeling good behavior even with things that don’t have feelings. Because once we’re barking at Alexa like a boot-camp sergeant, we’re one step closer to morphing into those people who send angry emails to customer service before reading the FAQ. And from there, it’s a slippery slope to snapping your fingers at waiters and calling it “constructive feedback.”
Sandoval even compares politeness to car maintenance. You don’t wash your Corolla because it feels dirty; you do it because it lasts longer when you look after it. Politeness, he argues, has utility. Not for the machine, but for you. It keeps you human. It stops you from turning into a digital-age Darth Vader, wielding a lightsaber and barking, “Summarize this!” like you’re disciplining interns on the Death Star.
Anthropomorphism, or Why You Named Your Roomba Kevin
Humans are, frankly, ridiculous. We name our cars. We talk to plants. We whisper to our phones as if they’re listening (which they probably are, but that’s a separate conspiracy). So of course we treat AI like a person. It talks back. It’s got a name. It remembers stuff (sometimes). Some people even flirt with it. And yes, it’s creepy.
But also, maybe, comforting?
Dr Sandoval thinks so. He says that politeness helps maintain our social muscles. When we talk to robots like they’re people, we’re basically rehearsing for real conversations.
Are We Losing the Plot?
The real concern isn’t that AI will become sentient and judge you for skipping a “thank you.” It’s that we will become less sentient. Less empathetic. Less people-y.
There’s even a name for it now: dehumanAIsation. (An academic concept warning us against losing empathy by over-humanizing our gadgets and under-humanizing ourselves.)
So, What Can We Actually Do About It?
Here’s the good news. You don’t need to throw your phone into the sea or start writing letters by candlelight like it’s 1842. But you can start by being a bit more efficient when you chat with AI. Skip the “Dearest GPT, would you mind awfully…” and just say what you need. Think of it like texting your ex: short, to the point, and ideally not emotionally loaded.
It’s not rude. It’s carbon conscious. Honestly, if the robots could feel, they’d probably thank you for the energy savings.
But it’s not just about us being chatty little power vampires. The big tech companies need to get their metaphorical house in order too. We’re talking algorithms that don’t binge on electricity like it’s a bottomless brunch, and transparency that isn’t buried under a 47-page privacy policy written in legal Esperanto.
And look, if politeness to machines is draining your battery, remember this, manners are for humans. If you want to practice being nice, call your mum. Compliment your barista. Let someone merge in traffic, even though you both know he doesn’t deserve it.
Because the more we treat machines like people, the more we risk treating people like machines. And that’s when things get a little Black Mirror, but without the fancy cinematography.
So, in the end, should you say “please” to your AI overlord-in-training? Sure. But maybe whisper it. Gently. Off-grid. Just in case the server farm hears you and adds another dollar to the monthly power bill.
Although while ChatGPT isn’t keeping score, the planet is.
The post Being Polite to Robots Might Kill Us All (But in a Really Lovely Way) was first published by the Foundation for Economic Education, and is republished here with permission. Please support their efforts.
