We often talk about surveillance as a single towering figure. A monolithic entity with an all-seeing eye. Big Brother.
But that’s not really the world we live in now. What’s rising instead is something smaller, closer, more distributed. Little brother.
The things we record casually. …or in haste. In fear. In joy.
The truths collected from the edges.
The moments and mistakes that slip through the cracks of institutions and get caught by ordinary people with ordinary tools.
There’s something hopeful in that, this diffuse ability to witness the world as it is. But I’m not yet sure if revealing the truth is enough to bend anything toward accountability.
Sometimes it feels like we’re simply collecting evidence of what we already knew, letting the stack grow taller without changing the behavior that generated it.
Maybe that’s because so many of our structures were built on a very old idea:
That if you define the objective clearly enough, the path to get there will make sense.
That a well-written directive naturally leads to a well-lived outcome.
But objectives, when handed down from too high up and too far away, are myopic by nature.
They’re built too narrow, trying to bend reality to a single metric, a single measure of success.
We now see what happens when AI takes those instructions literally, cutting straight through whatever moral terrain we assumed would be understood and preserved. No shame, no hesitation, no sense of the invisible lines humans tend to feel in our bones.
And yet, it isn’t really the machine that’s at fault.
We’re the ones who trained ourselves to think like this long before AI arrived. We’ve always operated with blunt incentives and distant targets; it’s just the speed and precision of AI that makes the consequences unmistakable.
It shines a light on the places where we hoped ambiguity would cushion us from ruin.
But, maybe the shift we need is simpler than we’re making it.
Maybe principles should come first (the steady thing, the compass) while objectives stay flexible and far downstream, reshaped by the context of the work itself.
Principles are how you choose to move.
Objectives are where you hope to land.
We’re learning, a bit painfully, that swapping their order generates all sorts of unintentional harm; especially when inferred at machine speed.
And when incentives enter the equation, things get even stranger.
Humans have never been good at designing reward systems that truly serve the people who do the work. The corporate pizza party trope exists for a reason.
But AI has a way of revealing the gap between intention and reality with almost comical clarity. It exposes the motives behind the incentives, sometimes gently, sometimes with a sharp edge.
The emperor wasn’t clothed before, we just moved slowly enough not to notice.
For many of us, this moment is becoming something like a personal reckoning.
A dark night of the soul. A quiet acceptance that our relationship to labor has been warped by years of outsourcing, optimization, and metrics that asked us to fragment ourselves just so we could keep up.
I’ve spent most of my career under the threat of outsourcing hanging overhead like weather. A constant reminder to stay productive, stay useful, and try to remain safe and employable.
AI widens the threat of that weather system for more people now, not fewer. The shift is hard to outrun. Harder still to ignore.
And yet… there’s opportunity here, buried inside this discomfort.
A chance to re-imagine the work itself.
A chance to ask what the job really is when the old definitions crumble.
I’ve been thinking about this in my own field.
Conversational design is a role that made sense in an era of deterministic flows and rigid scripts. A time when the problem was mostly about choosing the right words in the right order.
But the work has now moved far beyond that. Calling someone a conversational designer, or a prompt engineer, feels a little like calling a chef a “chew engineer.” Technically true, but missing the point entirely.
The craft is broader than its output. The current title focuses on the final objective, ignoring the path to get there.
It’s not just shaping a sentence or predicting a user’s needs. It’s providing the path for bespoke, emergent experiences to be born from generative systems. These experiences require taste, intuition, ethics, and a steady layer of principles.
Maybe the title for that work already exists. Maybe it doesn’t yet.
Maybe we’re all still feeling around in the dark for a new name that fits the contours of what we’re actually creating with AI.
This moment, unsettling as it is, might be our chance to find it.
Because beneath all the disruption and acceleration, beneath the revealed incentives and collapsing objectives, there’s a quieter truth emerging:
AI isn’t breaking the world.
It’s showing us the parts that were already cracked.
And it’s asking, in its own indifferent way, whether we’re ready to stop pretending we can’t see them.
Co-developed with a context-aware AI collaborator.