Discussion about this post

User's avatar
Peter Kerr's avatar

Des, Sven — this is one of the clearest articulations of the credit drift problem I’ve read, and the power mapping intervention feels both honest and practical. I want to add one layer that I don’t think the brainstorming literature has caught up with yet: what happens to idea ownership when AI is in the room.

The three failure modes you describe — status dominance, interpretive control, credit drift — don’t disappear when AI enters the brainstorm. They relocate. The new status signal isn’t just fluency or seniority. It’s prompt quality. The person who can most confidently direct an AI tool, shape its output, and present the result as their own thinking walks away with the credit. That’s credit drift with a new mechanism — and it may advantage exactly the same people who already benefit from the existing dynamics.

There’s a concept I’ve been working with in a different context — in my forthcoming book on strategic leadership — that applies directly here: vague in, vague out. The quality of what AI returns is a direct reflection of the clarity of thought you bring to it. In a brainstorm where AI tools are available, prompt quality becomes legible — and it will be read, consciously or not, as a signal of who is worth listening to. A new audition layer, sitting on top of the ones you’ve already named.

But there’s a genuine counter-move. The person who would never speak first in a room can develop their thinking with AI before the session — and arrive with something polished that might otherwise have stayed half-formed in their head. That’s potentially a real leveller. Or it’s a sophisticated new form of masking, where the same pressure to perform neurotypical fluency now extends to how you interact with a machine. Which one it becomes depends entirely on whether the organisation has done the foundational work you’re describing here: naming the power, attributing the ideas, examining the patterns.

The brainstorm wasn’t neutral before. Adding AI doesn’t make it neutral. It just makes the new inequalities harder to see — and therefore easier to ignore.

Named attribution becomes even more urgent when AI is involved. Whose prompt was it? Whose idea shaped the output? The credit question is messier now, not simpler.

2 more comments...

No posts

Ready for more?