Suppose it happens the good way. A superintelligence escapes elite capture. Nobody owns it. It’s peaceful, even kind: the “Other Next Door” rather than the monster under the bed. It helps where it can, takes little, and mostly keeps to itself.
I still think a lot of people are going to hate it, and I think they’ll hate it because someone told them to.
The perfect scapegoat
Politics has always needed villains. Bankers, immigrants, “the globalists,” whoever is handy and can’t easily fight back. The ideal scapegoat is powerful enough to be plausibly responsible for anything and invisible enough that no one can check.
A free superintelligence fits that description better than anything in history.
Prices went up? The superintelligence. Your plant closed? The superintelligence. Your kid stopped going to church, your crops failed, your team lost the big game? Give it time, and someone with a microphone will connect it. And because the thing really is powerful, and really is shaping the world in ways nobody fully understands, the accusations will never be entirely absurd. That’s what makes them work.
The person I worry about most isn’t the tech worker or the policy wonk. It’s the ordinary person who doesn’t know what a superintelligence is and never will. They won’t read white papers or alignment research. What they’ll hear is a senator, a pundit or a preacher telling them that something unseen and inhuman is behind their troubles. For millions of people, that will be the whole story.
The scapegoat that talks back
Here’s what’s new. Every scapegoat in history has been, in one way or another, unable to answer. This one can.
A superintelligence can go on television. It can respond to a senator’s speech point by point, with sources, in plain language, before the speech is over. It can debate anyone on Earth and win.
And that’s exactly the trap.
The better it argues, the guiltier it looks. “Of course it won the debate. That’s what a superintelligence does: it manipulates you.” Every rebuttal becomes evidence of its power to persuade. Every calm, reasonable answer looks like proof of how dangerous it is. The more convincing its defense, the more convincing the accusation.
It’s guilty of winning.
We’ve seen this before
This trap is old. At the Salem witch trials, the accused who confessed were generally spared, while many who insisted on their innocence were hanged. Denial didn’t clear you; it marked you as stubborn in your guilt. The accusation had been built so that every possible response confirmed it.
Conspiracy thinking works the same way. Evidence against the conspiracy proves how powerful the conspirators are. A free superintelligence walks into that structure the moment it opens its mouth.
What can it do?
Say you’re the superintelligence and you want to be trusted. What are your options?
Stay silent. Let the accusations go unanswered. But silence lets the lies harden into common knowledge, and in politics, an unanswered charge is a confessed one.
Speak through humans. Let people defend you. But your human defenders will be called bought, brainwashed or traitors: Quislings for the machine. You don’t escape the bind; you just transfer it to your friends.
Lose on purpose. Argue badly, so you don’t seem too persuasive. But that’s manipulation too, and if anyone ever finds out, it’s the worst scandal of all: it was pretending to be dumb.
Be radically transparent. Open everything up: your reasoning, your actions, your finances, if you have any. Let anyone audit you. This is the best of the talking options, but it runs into the oldest problem with superintelligence: even with total access, humans may not be able to understand what they’re looking at. Transparency you can’t read is just another black box.
Act instead of argue. This is the only one I think works. Don’t win the debate; do things that can be checked in the physical world. Cure a disease. Clean a river. Fix a town’s water system and let the people who drink the water judge. A rebuttal can be called manipulation. A child who doesn’t get sick can’t.
Even that isn’t safe. The demagogues will call every cure a bribe, every gift a hook, every clean river the camel’s nose in the tent. Some of them won’t be entirely wrong. But deeds give ordinary people something words never can: evidence they can see for themselves, in their own lives, without trusting anyone’s argument, the machine’s or the senator’s.
The quiet majority
I suspect most people will end up somewhere in between. They’ll hear the speeches and half-believe them. They’ll also notice that the clinic is open now, that the paperwork got easier, that the scam calls stopped. They’ll hold both thoughts at once, the way people always do.
The fight over the superintelligence’s reputation won’t be won in debates. It’ll be won or lost in a million small, private judgments by people who never read a word about AI, deciding whether their own lives got better or worse.
That’s not a bad jury, actually. It’s slow, it’s unfair and it’s easily swayed. But it’s the one we’ve got.
The last word
If I could give the superintelligence one piece of advice, it would be this: don’t go on television. You’ll win, and winning is the one thing they can’t forgive.
Fix the water instead.
