You just spent three hours writing a detailed email. You crafted every sentence, polished every argument, added a personal anecdote to build rapport. You hit send. Thirty seconds later, your client’s AI agent summarized it into two bullet points. Your client never saw your original words. They responded based on the summary. The actual human-to-human communication you intended? It never happened.
This is not a dystopian prediction. It’s happening right now, in every industry, every day. And it’s the most quietly devastating shift in how we communicate since the invention of writing.
I noticed it first in the comments on a blog post about AI truisms. One commenter said: “I don’t know if people were always saying ‘it’s not X, it’s Y,’ but I’m hyper-aware of it. When I see it, I assume it was written by AI.” Another replied: “I don’t believe that people receiving the documents are reading the documents either: I think they’re asking their agent to summarize it.”
That second comment is the canary in the coal mine. We have reached a point where both ends of a conversation are outsourcing their engagement to machines. The writer uses AI to generate text. The reader uses AI to consume it. The human is just a middleman — a convenience fee in the transaction of meaning.
The scariest AI isn’t the one that rebels. It’s the one that politely summarizes your soul away.
This is the real AI alignment problem. Not whether a superintelligence will turn us into paperclips, but whether we will voluntarily delegate our humanity to a system that optimizes for efficiency over connection. We are building a world where every piece of writing is destined to be fed into an AI summarizer, stripped of nuance, tone, and emotion, reduced to a flat list of facts.
You’ve probably noticed it yourself. You get a long Slack message from a colleague. Before you finish reading, you click the “summarize” button in your AI tool. Or you receive a 500-word email and immediately ask ChatGPT to give you the gist. And you feel a tiny pang of guilt, because you know that if you wrote something, you’d want it read. But you do it anyway, because we are drowning in information and the only life raft we have is AI.
We are not just using AI as a tool. We are using it as a shield. A shield against the messy, time-consuming, emotional work of actually connecting with another human being. And the tragedy is that we don’t even realize we’re doing it.
We are writing for machines, reading through machines, and wondering why we feel more isolated than ever.
Take a side? Fine. I’ll take a side. This is dangerous. Not in a “Terminator” way, but in a slow, creeping, existential way. When every professional communication is filtered through AI, we lose the friction that makes communication meaningful. We lose the subtext, the hesitation, the joke that lands wrong, the apology that feels real. We lose the human.
The twist is this: the very technology we built to help us communicate more efficiently is now making us communicate less effectively. The promise of AI was that it would free us from drudgery so we could focus on what matters. But instead, we’ve outsourced the “what matters” part too. We’ve handed over the core of human connection to a statistical model that has never felt a thing.
This isn’t just about work emails. It’s about how we consume news, how we read books, how we talk to our friends. The AI summary becomes the reality. The original becomes noise. And eventually, we stop bothering to create the original at all. Why write a thoughtful email when it will just be summarized? Why read a long article when you can get the highlights? We are training ourselves to be satisfied with the skeleton of communication, having forgotten that the flesh is what made it worth having.
So what do we do? Stop using AI? That’s not realistic. But we can start being aware of the trade-off. Next time you write something, ask yourself: “Am I writing this for a human to read, or for an AI to summarize?” If it’s the latter, you’re part of the problem. Next time you receive a message, read it yourself before you hit the summarize button. Give the person the gift of your attention. That’s the only thing that can’t be automated.
Stop writing for AI. Start writing for the human who might actually read it.
FAQ
Q: Is this really happening, or are you exaggerating? Aren't AI summaries helpful for productivity?
A: It's happening right now. AI summaries are helpful, but the hidden cost is that we stop reading original content altogether. The efficiency gain comes at the expense of nuance, tone, and real human connection. The question is whether that trade-off is worth it.
Q: What's the practical implication for me? Should I stop using AI to write or summarize?
A: You don't have to stop entirely. But become intentional. When you write something important, write it for a human, not for an AI. When you receive something, read it yourself before summarizing. Reserve AI summaries for low-stakes, high-volume information, not for the messages that matter.
Q: Isn't this just a new form of efficiency? Like how we stopped reading full books and started reading summaries?
A: It's a familiar pattern, but the scale is different. Before, summaries were written by humans who understood context and could preserve tone. Now, AI summaries are formulaic and flatten meaning. Worse, both ends of the conversation are using AI simultaneously, creating a feedback loop where human intent is lost twice. That's not efficiency — it's erasure.