The Call Is Coming From Inside the House

Every few weeks another very serious person warns that AI will exterminate the human race within a decade. It makes a great headline. It is also silly.

Here is the version that should actually keep you up at night — because it is mundane, it is already happening, and it is largely our own fault: AI does not have to want anything. It just has to be handed to enough people who do, and then help them switch off your power, your water, your bank, and your hospital.

The real risk isn’t Skynet. It’s cybersecurity finally sending the invoice.

Nobody’s civilization runs on a robot with red eyes. It runs on grids, pipelines, payment rails, and hospital networks — and almost all of them sit on machines that cannot prove what they are, cannot prove who told them to act, and cannot prove a human ever agreed to anything. We wired the whole world together and skipped the part where any of it could tell truth from forgery.

The doomsday scenario isn’t a machine deciding we should die. It’s an outage at civilizational scale — the lights, the water, the money, and the ER, all having a very bad Tuesday at once.

AI didn’t open the hole. It removed the guardrail.

Here is what the old risk models got wrong. For twenty years, the thing quietly protecting critical infrastructure wasn’t good security — it was the scarcity of competent attackers. Pulling off a real attack on an industrial control system or a payment network took a nation-state’s team. So the spreadsheet said “low probability” and everyone moved on.

That assumption just died. AI collapses the skill floor. What used to require an elite crew now needs a motivated amateur and a good model — the uneducated hacker, leveled up overnight, and multiplying daily.

And this isn’t hypothetical. Colonial Pipeline. Change Healthcare. Hospitals turning ambulances away because their systems were locked. Every one of those was pulled off by slow, tired, human crews. Now picture the same thing done by tooling that never sleeps, doesn’t need to understand the target, and tries ten thousand variations a night. The attack surface didn’t change. The number of people who can reach it did.

And here’s the embarrassing part: we already own the locks.

This is the twist that should make every CISO wince. Every defense you need already shipped — and we turned it off because security is inconvenient.

• TPM — a hardware security chip in essentially every business machine for two decades. Mostly never switched on.
• TEE and secure elements — a vault inside nearly every phone and processor. Barely used by the apps that matter.
• Self-encrypting drives — hundreds of millions sold, most never enabled to full strength. The engineer who invented them did it to solve exactly this problem. We shipped the lock and left it open.
• Signing — code and messages that could carry proof of who authorized them, sent naked instead.
• Secure messaging, and trusted consent screens — the one thing that makes “yes” actually mean yes, skipped because consent flows are annoying to build.

So we ran the global economy on unmeasured, unsigned, unverified machines and unproven human intent — and then acted surprised that it can be faked.

The call is coming from inside the house. We installed the deadbolts years ago. We just never turned the key.

It’s not just good governance. It’s new governance.

Don’t mistake this for a compliance chore. Turning these things on isn’t about passing next year’s audit — it’s that the whole model of governance just changed under our feet.

Old governance was paperwork. Write a policy, file a report, keep a log, run an audit once a year, and trust that the log is telling the truth. It worked — barely — because it assumed two things: that the actors were humans, and that there was time to check the work afterward.

AI breaks both. The actor is now a machine moving at machine speed, and “we’ll review it later” is worthless when ten thousand actions happened while you read this sentence. You cannot audit your way out of that. Trust us, we logged it is over.

New governance is different in kind. It doesn’t describe good behavior after the fact — it proves it in the moment. The machine proves what it is before it is trusted to act. Every action carries a signature saying who authorized it. Consent is captured as it happens, not assumed. The record is signed, not asserted. Governance stops being a binder on a shelf and becomes a property of every single transaction — which is the only kind of governance that survives when the thing being governed is a machine.

That’s the shift. Not better paperwork. Proof instead of promise.

This is not an apocalypse. It’s a to-do list.

Which is why this is the hopeful story, not the scary one. The doomers want you to believe a godlike intelligence is coming and there is nothing you can do — false, and paralyzing. The truth is more embarrassing and far more useful: you are exposed because you left the locks off, and the fix is engineering you already own.

Step one costs a hundred dollars.

New governance — signing, attestation, verified consent — all depends on one thing that has to come first: a name. You cannot sign as someone until you are someone a machine can verify. You cannot prove a message came from your company until your company has a verifiable digital identity for it to come from.

So before the expensive projects, before the platform decisions: every enterprise needs a digital name. It is the root everything else hangs off — the address the house needs before any deadbolt means anything.

It costs about a hundred dollars and takes a few minutes. Set up a DBA — a verifiable digital name for your company — at corpid.rootz.global. That is step one: the cheapest, most important thing on the list, and the first act of governing the new way.

Turn the key.

AI isn’t going to end the world. But it will absolutely find every door you left unlocked — faster, cheaper, and more often than anything before it.

The good news is you’ve been holding the keys the whole time. Start with the name. Then turn the locks on, one by one.

The house is fine. We just have to start acting like we live in it.