The UN Admitted Nobody’s Built The AI Answer.

The UN just admitted nobody’s built the answer to a question they can’t stop asking

There was a summit in Geneva this month. First one of its kind — the United Nations’ first ever gathering built entirely around AI governance. Government leaders, tech companies, researchers, all in one room, asking the same question out loud.

Who is legally responsible when AI causes harm?

That’s not a small question and it’s not new. But this time it had numbers attached to it. Experts told the summit that AI-generated harms are already showing up as human rights violations across multiple countries, with mounting evidence that’s getting harder to dismiss as theoretical. This isn’t a room debating something that might happen someday. It’s a room admitting something is already happening, and nobody in the room has an answer for who owns it. Aigovernance

Sit with that for a second. The United Nations, with every resource a body like that can call on, opened its first major AI governance summit and the headline finding wasn’t a solution. It was the size of the gap.

Here’s what makes this worth stopping on.

This isn’t a story about AI being dangerous. Everybody already knows AI can cause harm — that’s not news anymore. The story is that the world’s largest coordinating body just spent days in a room and came out the other side still asking the opening question. Not answering it. Asking it.

That tells you something about where the actual failure sits. It’s not in the harm itself. It’s in the fact that responsibility for the harm was never assigned before the harm happened. Nobody built the accountability structure first and the technology second. It went the other way around, the way it almost always does, and now a room full of the most capable people on earth is trying to retrofit blame onto something that’s already loose in the world.

That’s the gap The Faust Baseline was built to close, and it’s not a new argument here — it’s the same one, said louder by a bigger room this time.

CIMRP-1, the Moral Domain Protocol, carries the rule that harm doesn’t stop at the edge of a conversation. A person can be affected by a decision they never saw made, by a system they never interacted with directly. That’s exactly the shape of the problem the UN just described. The people getting hurt aren’t in the room asking who’s responsible. They’re the reason the question exists.

BLP-2 sits right underneath that. A system running honest boundary-naming doesn’t wait for a UN summit to figure out who’s accountable for what it does. It names its own limits before it acts past them. Every time. Not because a law requires it yet — most places, no law does — but because that’s the standard the conduct holds itself to, with or without a room full of diplomats catching up.

Here’s the plain truth under all of it. Accountability that gets built after the harm is not accountability. It’s cleanup. Real accountability gets decided before the system ever touches a person, not after the person’s already been affected and the world convenes to ask who’s to blame.

That’s the difference between a framework and a summit. A summit responds. A framework decides in advance.

The UN didn’t fail by holding this meeting. Holding it was the right move, and admitting the gap instead of pretending it’s smaller than it is took a kind of honesty most institutions don’t manage. But the meeting itself is proof of the exact problem it was called to solve — the world built the technology first and is only now, after real harm, trying to figure out who answers for it.

A protocol that names its own limits before it acts doesn’t need a summit to tell it where the line is. It already knows, because naming the line was the job from the start, not an afterthought bolted on once the damage was already done.

That’s not a policy position. That’s just the order operations have to happen in, if the goal is actually protecting people instead of explaining, later, why nobody could.