We’ve all received that email. The one that opens with “Per my last email” (translation: read more carefully next time). The one-word reply, “Noted.”, that somehow manages to feel like a door slammed shut. The wall of demands with no greeting, no “please,” no human warmth anywhere in sight.
It turns out this isn’t just a personal grievance. It’s a measurable workplace problem. And it raises a question worth sitting with: as we hand more and more of our correspondence to AI, are we about to automate one of our worst communication habits?
The quiet damage of a rude email
Email rudeness sounds like a minor annoyance, the office equivalent of someone stealing your stapler. But the research paints a heavier picture.
Some studies suggest as many as a third of employees receive at least one uncivil email a day. And those emails don’t just sting in the moment. Researchers have linked them to “work rumination,” that exhausting habit of replaying an exchange in your head long after you’ve closed your laptop. It makes switching off harder, and it’s been associated with higher rates of anxiety and depression.
Part of what makes email uniquely corrosive is that you can read it again. And again. A sharp comment in a hallway evaporates; a sharp email sits in your inbox, ready to be re-examined at 11pm. Academics studying this have even tied email incivility to physical symptoms like headaches, fatigue and disrupted sleep, and to knock-on effects at work, from absenteeism to eroded trust in managers. One line of research found that the stress doesn’t even stay at the office: it can follow people home and affect their partners over the weekend.
So why do otherwise decent people keep firing off these messages?
Why we’re rude on email (when we wouldn’t be in person)
A lot of it comes down to the medium itself. Email creates psychological distance. There’s no face on the other side of the screen, no flicker of hurt, no awkward pause, so it’s easy to forget there’s a real person reading. Researchers have compared it to how people behave in traffic: we’ll honk and gesture at a stranger in another car in a way we’d never dream of doing while standing next to them in a supermarket queue.
Add pressure, heavy workloads, fatigue, and the occasional bad mood, and the conditions are perfect for a curt reply. Often the sender isn’t even trying to be rude. Strip away tone of voice and body language, and a fast, efficient message (“Done.” “Fine.” “See below.”) reads as cold, even hostile. The reader fills in the emotional blanks, and they rarely fill them in generously.
Which brings us to the machines.
Enter AI: a fix, or more of the same?
Here’s the genuinely interesting part. AI writing assistants are now drafting a huge share of our work emails. In theory, this is great news for civility. An AI doesn’t get tired, doesn’t have a bad commute, doesn’t take a technicality personally and decide to make your afternoon worse. Ask it to soften a message and it will happily sprinkle in the “thank you so much” and the “I really appreciate your patience” that a stressed human might skip.
But there’s a catch. A few, actually.
AI learned from us. These systems were trained on enormous amounts of human writing, including all our clipped, passive-aggressive, “as previously stated” tendencies. If we’re not careful about how we prompt them, AI can faithfully reproduce the very tone we’re trying to escape. Ask for a “firm” or “direct” email and you may get something that lands a lot harder than you intended, now with perfect grammar and zero hesitation.
AI widens the distance. If email already creates distance between sender and reader, an AI intermediary stretches it further. When you didn’t write the words yourself, it’s even easier to forget there’s a person on the receiving end. The message becomes a task to dispatch rather than a human to address.
Politeness isn’t the same as warmth. AI is excellent at the form of courtesy: the greetings, the sign-offs, the softening phrases. But recipients are surprisingly good at sensing when warmth is hollow. A flawlessly polite email that clearly came from a machine can read as more impersonal, not less. We may be solving for “please” and “thank you” while quietly losing the thing that actually makes communication feel human.
So where does that leave us?
Maybe the honest answer is that AI is a mirror, not a cure. It will be exactly as courteous, or as cold, as we ask it to be, and it inherits our blind spots along with our vocabulary.
The fix, if there is one, probably isn’t technological. It’s the same one it’s always been: remembering that there’s a person on the other end. Reading the message back as if you were receiving it. Asking whether “Noted.” is really all this human being deserves.
AI can help us write kinder emails. But it can’t decide to be kind for us. That part, for now at least, is still ours.
Got a “per my last email” story of your own? We’d love to hear how email, and now AI, has changed the way your team talks to each other.
