The Hidden Cost of AI-Generated Communication

I've been thinking a lot lately about why I cringe when I receive a message that's clearly been run through an LLM for "polishing." It's not just a vague discomfort—I think I've finally figured out what's really happening when someone uses AI to rewrite their words.

The Problem With Genericization

When you send a message through an LLM, you're not just cleaning up grammar or improving flow. You're fundamentally altering the signal. We choose our words deliberately, even when we make mistakes. Those imperfect phrasings, those awkward constructions—they're part of the message itself.

But here's what's worse: you're also stripping away the context that lets me understand you. Every person develops a unique communication fingerprint. When you say "...we need to talk," I interpret that differently depending on whether it's from my partner, my boss, or my childhood friend. The words carry different emotional weights, different histories, different implications.

Building the Atlas of Understanding

Human communication works because we build an atlas of implicit knowledge about each other over time. Your writing style, your typical phrasing, your patterns of emphasis or omission—these become data points I use to decode your messages. When you run your text through a genericizer, you're breaking that system.

It's like trying to synchronize two watches by running one through a random number generator first. The handshake fails. The fabric that allows us to communicate effectively and honestly gets torn.

The Social Cost

What we're really losing is the opportunity to get to know each other. When you use AI to smooth out your rough edges, you're denying me the chance to understand how you actually think and feel. Those idioms that don't quite work in English? The moments when you're too direct or too flowery? That's you being authentic.

Make mistakes. Use the wrong words sometimes. Be imperfectly yourself. Because that imperfection is what allows real human connection to form. It's what lets me tune into your frequency rather than some generic, AI-generated approximation of human communication.

The Alternative

The next time you're tempted to run your message through an LLM, ask yourself: am I polishing this for clarity, or am I hiding behind a machine? Am I making it easier to read, or am I making it harder to understand the real me?

Give people the courtesy of interpreting your message in the context of who you actually are, not some sanitized, AI-optimized version. Let them get to know the real you, mistakes and all. That's where genuine communication—and genuine relationships—actually begin.

Original article by Sebastian Aigner