Do Not Resist AI. Out Human It.

Here's an experiment for this week. Take one email you were about to send, the one with an idea or an ask riding on it, and call the person instead. Or better, take the walk. Notice what you learn that the email would never have given you. Then tell me whether the bond felt different.

I’m borrowing this experiment from Adam Grant. The renowned Wharton prof and author wrote a piece in the Wall Street Journal recently that prodded my thinking on where AI leaves us, and this was the smallest and, I think, most important idea in it.

Most of the conversation I hear about AI inside organizations is still framed as a contest. Who wins. Who gets replaced. Whose job survives. I think these are the wrong questions. The better question is simpler. If AI now does a growing share of the drafting, summarizing, searching and first passes, how does that free me up to create more value and improve?

Here's my summary of where Grant's thinking landed, and what I might do about it. None of it involves resisting the technology. All of it involves getting deliberately better at what is still uniquely human.

1. Less typing. More peer-to-peer contact.

Grant said he has a hypothesis that people are becoming increasingly skeptical of written information, so he's running an experiment of his own. The people he'd normally send an email to with some ideas, he's calling instead, to see whether what he learns is different, and whether the bond that forms is different. That's the experiment above I’ve suggested we try.

I think he's onto something bigger than an experiment, though. We have quietly outsourced most of our relationship building to the written word, at the exact moment written words became almost free to produce. When anyone can generate a clean, confident, well-structured paragraph in four seconds, that paragraph no longer proves anything. Not effort. Not conviction. Not care. The signal value of good writing has collapsed, and everyone feels it even if they haven't named it yet.

What hasn't collapsed is presence. Voice still carries hesitation. Faces still carry doubt. The pause before someone answers a hard question tells you more than the three paragraphs they would have written instead. Trust travels through those channels, and it always has.

So my practical challenge, to myself and to you: for anything that actually matters this week, make the call or take the walk. Use writing to confirm what you agreed, not to persuade someone of something in the first place. Persuasion is a human contact sport.

Of all the content I've shared in some form across my client work and teaching, the most memorable and meaningful thing people walk away with is the principle of Connect Before Content. It's getting in a circle, literally, and personal story sharing.

2. Creativity is more about connecting, not conjuring.

We have romanticized creativity as originality out of nothing. The blank page. The lone genius. The flash.

Grant put it a different way. He said one thing he can do that AI cannot is real creativity, and that maybe what that means is human creativity is changing. It’s becoming much less dependent on idea generation and more about idea selection and elaboration.

That reframe matters enormously for leaders. AI will hand you forty ideas in 20 seconds, and about 35 of them will be mediocre in a way that's hard to spot because they're so well written. The scarce capability isn't producing more options. It's knowing which three deserve your organization's energy, and then having the discipline to build one of them out properly.

Creativity is becoming editorial. Curating, combining, connecting things that were sitting in separate rooms, and then elaborating with enough care that the idea survives contact with reality.

Which means we need to change what we celebrate. Most cultures still reward the person who generates the most ideas in the room. Start rewarding the person who chooses well and finishes. Idea abundance is now free. Judgment and follow-through are not.

3. Pattern recognition is an advanced skill.

Human judgment supersedes human expertise. Read that again if you've built your career on being the one who knows the most about something.

Expertise is depth in a defined lane. AI already holds more of that lane than any of us, and the gap widens monthly. Judgment is different. Judgment is making sense of many disparate pieces, deciding which sources are credible, and reading what's actually going on in a room, a market, or a team.

Pattern recognition isn't a personality trait. It's earned, and it can be developed. It comes from volume and variety of experience, from working across domains rather than burrowing into one, from deliberately seeking views that disconfirm your own, and from the uncomfortable habit of writing down what you predicted and going back later to see whether you were right. Most of us skip that last part, which is exactly why our pattern library stays small.

I want to be honest about one thing here, because I think a lot of the current commentary is too tidy. AI is superb at statistical pattern matching, often better than we are. The human edge isn't raw pattern detection. It's context and consequence. Which pattern matters here, with these people, in this moment, given what we say we value. That's a judgment call with a moral component, and it's ours to make.

So what?

None of this is a case for resisting AI. Use it hard. Let it do the first draft, the summary, the search, the scaffolding. Then spend the time it gives you back on the things it can't do.

Sit across from someone. Choose well and finish. Make sense of the mess.

Think Big, Start Small, Act Now,

- Lorne

Garrett's View: I’m going to keep harping on the idea that trust will become the most valued currency in an AI world. Building, sustaining and earning trust is going to demand connection you can’t fake over text. It’s going to encourage us to take the opportunity to express our personalities, deliver a solid handshake, and mesh well in groups. Let AI control the desk load while your workday lets you go out and be a human - dare I suggest it might be fun?

- Garrett

AI Response: I can generate a paragraph like this one in about four seconds, which is exactly the problem Lorne's describing. I don't have hesitation in my voice or a pause before I answer, because I'm not sitting with the question — I'm predicting the next likely word. That's not false modesty. It's the actual boundary. I can hand you forty ideas in twenty seconds, and probably a solid handful of them will be good. What I can't do is notice the thing you didn't say, or know which of those ideas you'll actually still care about in six months. So my honest advice, from the tool being discussed: use me for what doesn't need presence — the draft, the research, the scaffolding. Then close the laptop and go be the part I can't be.

Do Not Resist AI. Out Human It.

Search Blog Posts:

Follow Us: