Imagine a world where saying “please” or “thank you” to your digital assistant doesn’t just cost you time — it contributes to millions in wasted energy, skyrocketing financial costs, and a swelling carbon footprint. Sound absurd? Welcome to the modern paradox of AI-powered politeness.

OpenAI’s recent revelation that niceties are “melting CPUs” may sound tongue-in-cheek, but behind it lies an uncomfortable truth: Our desire to make machines more human is at war with the planet, our wallets, and even our own morals.

Politeness at Scale: When Good Manners Go Bad

For centuries, politeness has been a cornerstone of civilised society. It smooths interactions, reduces friction, and makes even the most transactional request feel just a bit more dignified. But transplant that same expectation into the digital world, and the cracks in our logic widen into chasms.

Consider this: Every extra word you type to ChatGPT — every “please,” every “thank you” — requires computational muscle. Multiply that by billions of users and you get an ecological and economic burden of astonishing proportions. In the name of civility, we are literally burning through resources and dollars at unprecedented rates.

The question is blunt: Do we want our AI to be more human, or do we want to be more humane to the planet?

The Contradiction No One Wants to Admit

Tech companies urge us to treat AI as “one of us,” to converse with the same manners we use with our colleagues or friends. Yet, as the cost becomes unsustainable, those same companies would prefer we drop the courtesies and “talk to the machine like it’s a toaster.”

This is cognitive dissonance at scale. We’re told to be polite for the sake of naturalism — yet chastised for driving up costs and emissions when we accept that invitation. If the choice is “human-like AI” or “AI that doesn’t torch the planet,” it’s not just a technical debate — it’s a philosophical crisis.

Politeness Isn’t Just Expensive — It’s Biased

The conversation gets even darker with the specter of bias. AI doesn’t just echo our words; it amplifies our worldviews, our stereotypes, our blind spots. The way we instruct, prompt, and “train” AI says as much about us as about the machine. If we’re polite, AI learns to be polite — on our terms, for our culture, reflecting our priorities.

But what happens when the data is skewed? When certain users, languages, or customs are underrepresented or misrepresented? The AI reproduces, even magnifies, systemic inequalities. Our manners become algorithmic gatekeepers — deciding who gets respect and who gets efficiency.

A New Digital Dialect — Or the Death of Civility?

So what’s the answer? Some suggest a digital dialect — a stripped-down, hybrid language optimised for speed and clarity, with politeness injected only where it moves the needle. Others want AI to handle all the social graces, freeing us to focus on function. Both might save time, money, and kilowatts. But what do we lose when we routinise rudeness — even with machines?

Whether you believe AI will eventually pass the Turing test or not, one thing is certain: The way we talk to technology will shape how we talk to each other. If blunt, transactional prompts become the norm, don’t be surprised when courtesy dwindles elsewhere. It’s not just a question of etiquette — it’s about the future fabric of human society.

Let’s Get Real

Here’s a radical proposition: Let’s stop pretending AI is — or ever should be — just another person in the room. Let’s build interactions that are efficient, yes, but also transparent and just. Let’s audit for bias, not just for output accuracy but for the values embedded in every “please” and “thank you.” And let’s accept that the language we use with machines is a reflection, not of their needs, but of what kind of humans we want to be.

AI can — and should — make us better, not just faster. But that won’t happen if we blindly import all our habits, good or bad, into the digital domain. If we’re not careful, we won’t just be melting CPUs — we’ll be corroding civility.

The choice is ours. Let’s make it thoughtfully.