The Wrong Question First

Why the AI Debate Has a Deadline Problem

Start with three real questions about AI. Not the only three — the three with a paper trail. The rest can wait. Not because they don't matter. Because they come after the foundation, not before it.

What does it cost the planet to run.

Who owns what it produces — and who owned the material that trained it to produce anything at all.

Who is accountable when the data inside it leaks, or is stolen, or is used against the people it came from.

These are not abstract questions. They have numbers. They have precedents. They have people who could, right now, be held to answer for them.

You don’t open a business without a guarantor. You don’t build a house without a title, an owner, someone accountable if it collapses. Every serious undertaking starts here — ownership, accountability, who answers if it fails — before anyone asks what the thing means.

This is the one place we’ve seemingly run it backwards.

Nobody debates a house’s personality before laying its foundation. We start with the parts that determine whether the house can be lived in at all.

None of them require us to agree on whether a machine can feel anything.

And yet the debate that dominates every panel, every op-ed, every dinner table argument, is not any of these three. It’s whether the tool is conscious.

That’s not a more important question. It’s a more convenient one.

You’ve seen this move before.

Not in tech. In a much older argument, one that gets rehearsed every time an industry needs a question to stay open a little longer.

For decades, the question of whether cigarettes caused cancer was treated as unsettled — not because the science was unclear, but because staying unsettled was profitable. The industry didn’t have to win that argument. It just had to keep it running. Every year the question stayed open was a year nothing had to change.

This debate has done something more elegant. It didn’t just keep a question open. It gave the object an inner life to argue about. Now the debate isn’t “is the tool dangerous” — it’s “does the tool have feelings.” That’s not a deflection anymore. It’s a philosophy department’s worth of deflection, with tenure. And while the seats argue, the data centers get built, the water gets used, and the people who need the tool most are the first to lose access to it.

Here’s why it wins.

Environmental cost has a number. You can measure it, audit it, compare it year over year. Someone can be held to that number.

Ownership has a paper trail. Contracts, training data, terms of service. Contested, sometimes hidden — but not infinite. A court can eventually rule on it.

Security has a breach date. A record of what leaked, when, and who it belonged to. There is always, eventually, an incident report.

Even “will it replace us” has a shape. Labour statistics. Layoff numbers. An economy that will tell you, in real time, whether the fear was justified.

Consciousness has none of that. No instrument measures it. No court can rule on it. No incident report will ever say: confirmed, it felt something, here is the timestamp. It’s the only question on the list that never resolves.

Whether intentional or not, the effect is the same.

A question with a deadline eventually forces someone to answer, or be seen refusing to. A question with no deadline just runs — forever, for anyone in the room, at any panel, in any op-ed — and running is enough. Nobody has to lose the argument. Nobody has to win it either. The room stays occupied, and occupied is exactly as useful as answered, if what you need is time.

That’s the vacuum. Not silence. Occupancy.

So who benefits from a room that never closes?

Not the planet paying for the compute. Not the people whose data trained the thing without asking. Not the people whose information sits inside it, one breach away from being someone else’s problem. Every one of those questions has an answer waiting, and every one of those answers costs somebody something — a fine, a policy change, a redesign, a number that has to go down instead of up.

Studying consciousness is expensive. Debating who would be accountable for it is free. It can be discussed at every conference, in every op-ed, on every panel, for years, without a single dollar changing hands or a single policy shifting.

The Swedes have a phrase for this: spel för kulisserna. Theater for the backdrop. A performance staged for how it looks from the seats, with nothing happening behind the curtain that the seats will ever see.

And when the public does manage to ask a pointed question, watch where it lands.

Not “who profits.” Not “who’s liable.” Not “who owns the data going in and the outcome coming out.”

It lands on you.

AI is making your brain work slower. AI is making you lazy. AI is killing your critical thinking. Every version of this argument does the same thing — it takes a question that should be aimed at the builders, the owners, the ones cashing out, and quietly re-aims it at the person holding the tool.

The hand that built the tool stays out of frame. The hand that pointed it stays out of frame. The person the tool was used on becomes the one under investigation.

That’s not the vacuum staying silent. That’s the vacuum producing noise that sounds like accountability without costing the people who built the thing anything at all.

The next time someone asks whether AI is conscious, notice what isn’t being asked in the same breath.

Notice what it cost to run. Notice who owns what it made. Notice who’s exposed if the data underneath it walks out the door. Notice who gets blamed for using it, and who never gets asked what they built.

None of that requires a verdict on whether the tool has an inner life. It requires someone to answer for the outer one — the one with a power bill, a contract, a breach date, a user taking the fall for a system they didn’t design.

Ask yourself, next time a debate feels endless: is it endless because it’s unanswerable — or because answering it would cost someone something?

The question of whether it feels something can wait. It’s the only question in the room that can afford to.

Everything else has been waiting long enough.