It starts with a Slack message. Or a PR comment. Or a doc link in Notion. Something about it feels off. Maybe it says, “In today’s fast-paced work environment…” and you feel your soul slip out the back of your head. The words are spelled correctly, the grammar checks out, the formatting is even passable. But it reads like toast that’s been left out a little too long: technically edible, but dry, stiff, and mildly exhausting.

Welcome to the era of lazy AI-generated work.

It’s not that AI output is inherently bad. In fact, it’s often too good: grammatically polished, structurally sound, and eerily confident. But it lacks intention. It doesn’t know what matters to you, the reader. It doesn’t know your team, your tech debt, your context. And when someone slaps that kind of text into a work conversation, unedited and uncurated, it sends a clear signal: I didn’t really care who I was talking to.

Disengaged communication #

There was a brief, shining moment when AI looked like it might elevate our work. Draft faster! Express ideas more clearly! Summarize long threads! Instead, what we got were unreviewed, bloated pull request descriptions, Slack messages that read like press releases, and strategy docs indistinguishable from LinkedIn posts. And here’s where the friction sets in. When you paste AI text without tailoring it, you’re not just saving time. You’re asking someone else to spend more. More time reading. More energy parsing what’s signal vs. boilerplate. More work guessing what you actually think. This isn’t a workflow enhancement. It’s a work redistribution.

Trust, Judgment, and Attention #

Human communication carries implicit contracts. Sending a doc suggests I considered what you’d care about. Leaving a code review comment signals I read your change and applied judgment. These acts imply effort, attention, and a basic level of respect.

AI, when used carelessly, breaks those contracts. It pads weak ideas with confident phrasing, substitutes volume for clarity, and mimics thought without understanding. It’s all surface, no spine. An architectural decision memo that says a lot but means very little. If your doc makes me wonder whether you even read it before sending it, you’ve already lost me. Trust is slow to build and fast to evaporate, especially when the message feels more like output than intention.

Synthetics in the Wild #

Let’s call out a few common patterns. Ending a Slack message with “In conclusion…” as if you’re defending a doctoral thesis. Writing code comments that invoke phrases like “deliberate design choice” but can’t explain what was deliberate or why. Dropping in docs that cite best practices without a shred of relevance to internal constraints. Or adopting a tone that sounds like an HR handbook eloped with a TED Talk. Toss in a few too many em-dashes and a casual overuse of “it’s crucial,” and suddenly it’s less ‘actionable insight’ and more Mad Libs for middle managers. These aren’t crimes, of course. But they are symptoms of disengagement. And over time, they make teams worse: slower, less trusting, and more inclined to quietly ignore each other’s work.

Use the Tool, Don’t Be the Tool #

To be clear: I use AI. It’s helpful for outlining, rubber-ducking, and getting past blank page paralysis. But there’s a difference between using AI to support your thinking and outsourcing your thinking entirely. Treat AI drafts like input from a junior teammate. Edit aggressively. Add context. Why does this matter now? Why are we talking about this? Own the voice. Inject your tone, your judgment, your caveats. And if you’re pasting something mostly untouched, at least flag it as such. Let others decide how much trust to allocate.

Before you hit send on that wall of text, try a quick gut check. If someone sent this to me, would I read it? Would I trust it? Would I feel like the sender actually thought about me? If the answer is no, it’s not ready. Even if it passes Grammarly.

AI can summarize, expand, reword, and codify. But it can’t care. That part’s still on you. So do your reader a favor: write like you’re asking for their attention, not demanding it. At a minimum, write like you respect their time. They might even respect yours back. Or at the very least, don’t let the robot get the last word.