I recently exchanged emails with a growth lead at a startup. His messages were clean, professional, and perfectly structured. I used AI to craft my replies—polished, persuasive, on point. For a few rounds, it felt like two well-oiled machines talking. Efficient. Clear. A little… hollow.
Then we hopped on a call.
Within minutes, the vibe shifted. We laughed at a clumsy joke. Heard the pause before a real answer. Felt the sincerity—or hesitation—in each other’s voice. It was human again.
That got me thinking.
Today, I came across a tweet about companies using AI to conduct early-stage interviews. My first reaction? Fair enough. If companies use AI to screen candidates, why shouldn’t candidates use AI to prep, polish, and maybe even respond?
But then the question deepened.
What if we extend this beyond interviews?
What if AI speaks for us not just in business negotiations, but in dating? In asking for a favor? In persuading a friend? In any delicate moment where we want to be convincing—but also real?
We’d optimize tone. Remove friction. Maximize persuasion.
But we’d also remove the stumbles, the vulnerability, the unscripted honesty that makes a connection meaningful.
I’m not against AI as a tool. It can help us articulate ideas, save time, and reduce miscommunication. But when both sides are optimized—when communication becomes AI talking to AI—what remains of the human in the exchange?
Efficiency at the cost of authenticity? Clarity at the expense of character?
In code, we refactor for performance. In communication, I wonder: are we optimizing away the very things that build trust?
So, I’ll leave it to you: Where should AI stop speaking for us?
Have you ever felt the “gap” between an AI-crafted message and a real human moment?
Top comments (5)
Your anecdote already shows the answer: the email did friction work (efficient, structured, moved things forward), while the call did resonance work (presence, pauses, vulnerability). Instead of asking where AI should stop, the better question is what kind of work is this moment doing?
• Friction work: scheduling, formatting, structuring arguments, reducing ambiguity—AI can carry this.
• Resonance work: clumsy jokes, hesitations, timing—trust lives here, and it needs rough edges.
The danger is mistaking friction‑handling for resonance‑creation. A polished condolence note may look perfect but feel like cardboard.
I love that you put words to those types of work: “friction” vs. “resonance.” The challenge, as you point out, is knowing which kind of work a situation is asking us to do. Polished AI messages handle friction well, but resonance still needs the human touch.
I can imagine a future where AI handles most (or even all) business communication, leaving little room for resonance; that feels plausible to me. And while your email example works well, it wasn’t quite a high-stakes business context (I think). My concern is that partnerships may stay efficient but never become relational, because there would be no space for them to grow into something more human.
If I already sense that this interaction will matter, I don’t want AI anywhere near it. Not because I distrust the tool, but because I trust the moment. Resonance needs authorship, not assistance.
Everyone is talking about Al. Al this. Al that. Al is taking over the world.
Which Al?
Al Pacino? Al Bundy? Al Capone? Al Yankovic? Big Al Delvecchio?
😂 AI Rick maybe