The Missing Variable
A study published last June by MIT Sloan found something that should matter to anyone trying to figure out how to use AI well.
Generative AI does enhance creativity in the workplace. But only for people who have strong metacognitive strategies and skills.
Not better prompts. Not more sophisticated tools. Metacognition. Thinking about your thinking.
The researchers ran a field experiment with 250 employees at a technology consulting firm. Randomly assigned people to either use ChatGPT or not. The ones with access to AI were rated as more creative - but only if they already had strong metacognitive habits. They could reflect on what information they lacked. Monitor how effective their approach was. Reassess when things weren’t working. Refine accordingly.
Everyone else? No significant boost.
Jackson Lu, the lead researcher, put it this way: “To fully unlock their creative potential, employees must know how to engage with AI—to drive the tool, rather than letting the tool drive them.”
If you’ve heard me talk about AI, I often say everyone has a Ferrari and no one knows how to drive. This is the same point. The tool isn’t the bottleneck. The driver is.
What Metacognition Actually Looks Like
This isn’t abstract. When it’s working, you notice yourself doing things like:
“I’m rushing here—let me slow down.”
“I think I’m protecting something. What is it?”
“This feels like I’m circling the same idea. What’s underneath it?”
You treat your own thinking as something you can look at. Work with. Adjust.
“Metacognition—thinking about your thinking—is the missing link between simply using AI and using it well. It allows people to ask better questions, recognize knowledge gaps, and extract real value from AI tools.”
—Jackson G. Lu Sloan School Career Development Associate Professor of Work and Organization Studies
And there’s another piece to it. When you say something and have that thinking reflected back to you in words, you either know it feels right or it doesn’t. That’s you exercising taste and judgment. Often there’s a feel. You feel it if it’s off. You really feel it if it hits.
When metacognition isn’t working, you stay abstract. You intellectualize. You frame instead of feel. You produce without noticing how you’re producing.
The difference matters because AI amplifies whatever you bring to it.
If you bring half-formed thinking and sit with it, the AI can help you find structure. If you bring premature closure, it’ll polish something that wasn’t ready to be polished.
This Connects to Work I’ve Done My Whole Career
I’ve spent years facilitating high-stakes work with senior leadership teams. UNICEF. The Red Cross. Doctors Without Borders. Large nonprofits and public organizations in Canada, the UK, and globally. Often working directly with executives on complex issues where conventional strategy breaks down and human dynamics determine whether anything actually moves.
And what is facilitation, really?
It didn’t matter what context I was in. It was always the same thing.
Holding space for a group. Creating conditions for people to think productively together. Making ideas visible—turn them into sticky notes, put them on walls—so we could actually work with them instead of talking past each other. We could see what we were talking about.
A sticky note was never just a sticky note. It was your idea made visible. Made real. Made into a thing we could move, combine, build on.
That’s metacognition at a collective level. Thinking about thinking, but together.
And what I’ve realized is that working with AI is the same skill set.
You’re creating conditions for productive thinking. You’re making ideas visible so you can work with them. You’re diverging—exploring, playing, seeing what’s there. Then converging—deciding, landing, clarifying what you actually think.
The exercises I’ve been sharing in the paid subscription are all versions of this. Daily reflections. Team conversations captured as transcripts. Personal check-ins about what’s working and what isn’t. Goal-setting that pulls from everything you’ve already explored.
All of it is making thinking visible so you can think about it.
The Real Point
Three years in, there’s been an intuitive way for me to approach this technology that has resulted in really great outcomes. And as I write, share what I’m learning, work a bit in the open—it’s nice when you come across something like this study that gives language to what you’ve been experiencing.
The technology itself is magical. Someone said on a podcast recently that this is the closest thing to real life magic we’ve ever had. And it kind of is.
But it isn’t actually magic. We don’t fully understand why it can do what it can do. But we do have ways to work with it. You can build this as a skill set. And it’s not about being a technical person. Not at all.
My friend, Dr. Linda Carson, always said this about creativity and collaboration: there are fewer rules than you think. She has passed away, but metacognition was a constant topic as we developed ways to teach people about creativity.
Working with AI can be an incredibly creative experience. That’s what I keep discovering. That’s what this study confirms.
AI doesn’t make you more creative. AI plus metacognition does.
The gap isn’t access to tools. Everyone has access now. The gap is the human capacity to use them well. To drive rather than be driven.
That’s learnable.