Johan Roos co-invented LEGO Serious Play — the method that has leaders think with their hands instead of their mouths, because the medium you use changes what you can see. Now, with everyone pouring every ounce of oxygen into AI, Johan has written a book arguing that the same forces can quietly erode the very human capabilities that method was built to unlock. And the kicker: he and Bill both have sons working at the leading edge of AI, so this isn't a view from the sidelines.
In this episode, Bill and Johan circle the heart of Johan's new book, Human Magic: Leading with Wisdom in an Age of Algorithms — and they start, fittingly, with the physical. Bill talks about wing foil surfing; Johan about riding horses and shoveling out the stable. Both describe the same thing: embodied, full-body experience that AI can't touch. From there they get into Johan's central claim — that every single day, in tiny micro-decisions, you either amplify or erode your own curiosity, creativity, critical thinking, communication, and collaboration. One path takes energy and sharpens you; the other is soothing, invisible, and slowly fills your work with the machine's output instead of your own. It's a choice, he insists, not a fate.
They get practical fast. Johan's first rule for leaders: if you're not using AI yourself, every hour of every day, stop talking about it. He tells the story of standing in front of 1,300 business school leaders, running an old-school Oxford Union debate and then a live AI voice assistant on stage — because the medium is the message. He lays out his "wisdom compass" — four questions (what matters, what is right, what resonates, what works) that should stay anchored in you, not outsourced to a large language model ("they're large language models, not large thinking models"). They talk sycophancy, the erosion of reasoning relative to feeling, Toulmin's argument mapping for critical thinking, and the danger of teenagers and young professionals using AI as a substitute for their own grunt work. Johan's distinction for companies: don't just redesign roles around AI (structural erosion) — also protect how your people think (behavioral erosion). His closing challenge to leaders: subtract the AI output from your last big discussion — what's left, and how do you measure it? It earns him blank faces every time.
In This Episode
Guest Links
Host Links
Mentioned in this episode:
PhD Research on CEOs
Quick favor: I'm in the middle of my doctoral research, and I need CEOs
Here's the question: is there a point where working more hours actually indicates a worse leader? Nobody has measured it. I'm measuring it.
If you're a CEO or president with at least ten people in the business and three direct reports, it takes just a few minutes. And Your team answers a few anonymous questions.
What you get back is a report comparing how you rate yourself to how your team rates you. That gap is usually the interesting part.
ScalingCoach.com/study
