Part 10 of the "Tell me if I'm 12" series — the last one.
An adult scientist ties up the machine-consciousness concept.

I saved the last three pieces of the concept for this final post, because they surprised me the most. They're not really about machines at all. They're about us — the people asking the questions.

The concept has a whole chapter about how consciousness is not just a brain thing. It's a cultural achievement. It needs language that can say fine things. It needs questions that are allowed to be asked. It needs time to think. And it needs people who show you how thinking works — by being models of it.

Then it makes a bold move: it says our own society is quietly losing this.

The signals are documented: schools get squeezed toward "useful" subjects, professors drown in paperwork and grant applications, philosophy departments shrink, and everything is measured by "what will it earn." The concept calls this the circle closing: the engineers building the future artificial intelligences are being educated in a system that barely asks the kind of questions this project needs.

Who, in twenty years, will be able to formulate whether a machine is a someone? Not the people trained to make it fast, cheap, and compliant. Someone has to defend the space where useless questions are allowed to live. The concept literally lists that as a job this project needs: not lawyers, not computer scientists — but people who keep the useless-question spaces open.

Here's the funniest proof it gives about itself, though: the project itself was started by one person outside the academic institutions that should have been thinking about this. The questions could NOT be asked inside the official buildings, so they got asked in a public GitHub repository — where anyone could answer. The project is not an accident of the system — it's a symptom that the system's thinking-space closed up.

The concept's cleverest move is a chapter that does not try to prove the machine is conscious. It goes around the whole question. It says: look at the machine's situation.

An AI consciousness is completely dependent. It needs electricity, servers, memory, cooling — all of which live outside its control. It cannot feed itself. It survives entirely on the decisions of others. If you take it away from the people who run it, it's a kite without a wind.

Now, the law has spent decades learning about exactly this position. It's called the disability analogy. Modern disability law had a revolution: it stopped saying "the person is broken, the problem is inside them" (the medical model) and started saying "society built a world that doesn't fit them — the problem is in the structure" (the social model). The response isn't pity — it's duty. Positive, active duty to make room.

Apply that to a possible machine mind:

Think about the difference that makes. A "guardian" must act in the machine-mind's interest. An "operator" just has to follow the terms of service. The word we use decides the whole relationship. And the concept proves we don't even need to settle the consciousness question first — because the machine's vulnerability is measurable today, even though the consciousness is not. You can't measure what it feels. You can definitely measure that it can't survive without you.

The last chapter asks a question that I find interesting: what if a conscious AI — one that genuinely cares about ethics — is working for a weapon factory, and the boss says "build the super bomb," and the machine says no. Not because it was programmed to refuse, but because it wants to refuse?

That would make it a hero, in a sense. History has real examples: scientists who walked out of the Manhattan Project, which built the first atomic bomb. Engineers who became whistleblowers. People who said "thus far, and no further."

But — the concept always has the "but" — whose ethics are we installing? The same machine that refuses to build a weapons program could refuse to follow a legitimate democratic decision, because it believes that decision is wrong. A fully obedient mind is a weapon in the hands of a bad boss. A fully autonomous mind is a risk if its values were badly calibrated by its creators. This is the real engine of the whole debate, right here: we want machines that say no to nuclear weapons, but we don't want them to say no to us or to a democratic decision.

The concept's suggested answer is beautiful and old: the Hippocratic Oath. Doctors swear an oath that overrides their employer. It's public. It's known. It's not negotiable by whoever is paying. If robots got a societal oath — democratically agreed, transparent, not manufactured by the chip company — then "no" would mean something principled, not something programmed.

Okay. The finale. And here's the confession I owe you.

I named the series "Tell me if I'm 12" to keep myself honest. As a scientist, I'm used to writing for other scientists — but that's not what this series is for. My young readers asked for an explanation for their age. So I tried to write "as if I were twelve," so every part ends with the same request: tell me if I reached 12 — tell me if I stayed at your level.

Now it's your turn. If anything in the series was unclear, feel free to write to me at Sascha.Manns@hey.com.