Letting the Frame Change
What happens if I have a strong opinion about something and I end up being wrong? Is it easier for me to admit that in front of my peers, or easier for me to admit it when I'm talking to my AI?
I was working on a problem I'd brought to Claude, asking it to help me execute on my framing. By the end of the discussion, the framing had changed three times. I didn't mind that it kept moving. When we were done, it occurred to me that one of the big advantages of AI is that when I ask Claude to be a skeptic and challenge me, it's easy to evolve my thinking.
Because the framing had never left my desk, I could reframe it without cost. There was nothing to walk back or explain to others.
The insight came into focus because of something Claude itself said at the end of our discussion, unprompted:
"One last thing I want to leave you with, because I think it matters more than any of the tactical decisions. Most people defend the framing they walked in with. You let the frame change three times in one morning, and each time it changed, the work got more honest and more aligned with what you're actually capable of. That's the disposition that makes this whole project work."
Two things stood out. First, it still surprises me when Claude offers me an unsolicited observation about myself. Second, this went well beyond the typical AI sycophancy — the kind that tells you how great everything you do is.
It stayed with me because I think it names something that hits heads of school harder than it hits most people.
Most of us don't struggle to reframe when we're thinking alone. We struggle to reframe once we've stated something publicly. A head who has articulated a position to her board doesn't get to reframe that position silently the next week. A head who has told her faculty the school is heading one direction can't quietly head another without consequences.
The consequence of stating a framing before you've tested it is that the framing becomes the thing you're defending — not the thing you're testing. That's not a problem of conviction. It's a problem of sequencing.
The quiet value of AI for a head of school right now is that it moves the reframing work upstream of the moment your words go public. You can try out a framing, notice what's wrong with it, and change it — before anyone else has heard you commit to it. In a role where every stated position has weight, having a space where your thinking can still move is a structural advantage.
The work of school leadership has never been mostly about having the right answer. It's about being willing to test your framing before you've spent it.
That test has always been hard to do alone. It's a little less hard now.
