The Hidden Cost of AI Isn’t Automation. It’s Authorship.
Decision-making under zero latency
My book, EDGE, took an unexpected turn.
I started writing it as a field guide to leadership in fast-moving environments. How to lead well in the Intelligence Age. How to make decisions when conditions keep shifting.
Partway through, it became clear that wasn’t the real problem.
The book needed to be about authorship.
The shift came from a concept in systems engineering called edge computing.
In technology, edge computing exists because systems became too fast for centralised decision-making to keep up. When data has to travel back to a distant cloud, decisions that should take milliseconds start taking seconds. By the time a response arrives, the moment that required action has already passed.
That delay is called latency.
So systems adapt. Intelligence moves closer to where data is generated and where consequences occur. Not because it’s elegant. Because delay becomes intolerable under acceleration.
What’s changed now is that human systems are under the same pressure.
Human perception operates in fractions of a second. We notice a shift in tone, a message, a signal, almost immediately.
AI systems now operate at similar speed. Outputs appear instantly. Actions trigger without pause.
Institutions do not.
Institutional decision-making still runs on minutes, hours, days. A signal appears. It’s logged. Reviewed. Escalated. Discussed. Deferred.
By the time a formal decision arrives, the moment that required judgement has already gone.
Nothing breaks. Everyone follows process.
But decisions don’t stop. They migrate.
They show up in Slack threads, shared documents, dashboards, prompts, defaults. Work keeps moving, but no one clearly decides what matters, what stops, or what consequence they are prepared to carry.
This is the authorship problem.
An author decides what belongs and what does not. They are accountable for coherence. When something fails, they cannot say, “I was just responding.”
A reader reacts to whatever appears next.
Under accelerating, tech-driven change, many capable people begin reading their working lives instead of authoring them.
They stay busy. They respond quickly. They deliver.
But nothing is clearly owned.
AI doesn’t remove human agency. It reshapes the conditions under which agency is exercised. Systems act by default. Outputs compound without friction. Decisions accumulate whether anyone claims them or not.
The risk isn’t automation.
The risk is decisions compounding without authorship, just as their consequences accelerate.
Over the quiet weeks of the holidays, I’ve been finalising the manuscript and preparing it for print. The first copies of EDGE: A Field Guide to Operating at the Edge of Tech-Driven Change will be available in late January.
Before it leaves my hands, I keep returning to one question:
Where would it be easier to defer judgement than to own it?
Because the question is no longer whether decisions are being made at the edge.
They already are.
The question is whether those decisions are authored, or defaulted.
That’s the edge I’m writing from.
To the edge and beyond. See you out there!
Kate