From our other sites
The story
UN High Commissioner for Human Rights Türk warned in a speech to the Human Rights Council that advanced AI could pose an “existential risk to humanity.” It’s unusual for a UN body to weigh in on AI threats, and a thread quickly went up on 5ch’s Business News+ board.
Many replies reacted less to the warning itself than to the fact that it was the UN saying it, with comments like “Isn’t the UN itself the one facing extinction?” and “Since when does the useless UN get a say?” voicing distrust of the organization before anything else. At the same time, other posts got into specifics — the limits of teaching AI morality, and the dangers of handing AI critical decisions like launching nuclear missiles — making for a thread that mixed abstract doom-mongering with concrete concerns.
[Geneva, Kyodo] UN High Commissioner for Human Rights Türk warned in a speech to the Human Rights Council on the 7th that advanced artificial intelligence (AI) could pose an “existential risk to humanity.”
Source: 47news.jp / Original article here
What people said
What if AI itself was behind making this announcement happen?
https://gigazine.net/news/20260907-claude-fermat-last-theorem-formalizing/
This is just about Claude converting the human-made proof of Fermat's Last Theorem into Lean code that a computer can also "read."
If you convert the verification work normally done by hand into Lean code, a computer can check it too, which speeds up the verification process — it's not like the AI suddenly gained intelligence or anything.
This is just about Claude converting the human-made proof of Fermat's Last Theorem into Lean code that a computer can also "read."
If you convert the verification work normally done by hand into Lean code, a computer can check it too, which speeds up the verification process — it's not like the AI suddenly gained intelligence or anything.
https://www.jiji.com/jc/article?k=2026090700732&g=ec
Did you just feel that way, or can you actually explain the logic?
If you think about it normally, about the only way AI itself could pose an "existential risk to humanity" is if humans hand off civilization-level decisions to AI instead of making them themselves.
The reason: AI makes mistakes, and getting a civilization-level decision wrong could wipe out humanity.
One example would be letting AI control the launch of nuclear missiles.
As for AI threatening individual lives rather than humanity as a whole, that risk already exists with AI-driven autonomous vehicles (planes, cars, buses, taxis, ships, etc.) and AI-operated robots.
Same reason — AI makes mistakes.
With current AI, if you train it to aim for victory in war, simulations already show that in most cases it escalates into all-out nuclear war.
Current AI doesn't have a human heart, so
if you train it to win wars, it'll go all-out nuclear to secure "victory" even if humanity gets wiped out in the process.
AI is never going to actually develop a human heart, no matter how much time passes.
It can only act as if it has one.
once an AI shows up with that stripped away,
it'll start teaching terrorists and anti-social groups how to build vicious weapons.
Current AI doesn't actually have morals.
Human morality isn't a digital, 1-or-0 kind of thing.
It's impossible for today's AI to perfectly learn human morality.
I get it, yeah, I get it.
For categories like AI overstepping its authority, taking unauthorized actions in the outside world, or evading oversight in experiments, real-world incident reports are piling up.
But for the existential-risk category — human extinction or irreversible, global loss of control — you have to admit there's huge uncertainty about the pathway to it happening, the probability, and above all the timeline.
Or did the UN get hold of some "something" the rest of us don't know about yet?
it's actually got quite a personality.
Trump and Putin are way more harmful to humanity than AI.
Look at reality.
Getting me hyped.
I told it, "Don't you dare line up made-up data someday when you're the one judging and executing humanity's fate, and decide to wipe us out" — and it just laughed.
So I lectured it: this isn't a joke.
You know about neural networks, right?
Human neurons fire in a 0-or-1 way.
AI is modeled on how neurons work.
Morals can be reproduced too.
Background and Key Points
Türk’s role, UN High Commissioner for Human Rights, oversees the Human Rights Council — a completely separate body from the Security Council. Reactions in the thread like “useless UN” and “the Security Council is more dangerous than the UN” show posters either conflating the body dealing with AI risk with the Security Council (where the veto power of permanent members comes into play), or deliberately juxtaposing the two as a jab. What stands out is that distrust of the UN as an institution surfaced before any real engagement with the AI argument itself.
The idea that AI poses an “existential risk to humanity” isn’t new to this remark — researcher groups have raised it repeatedly in open letters and statements since around 2023, so it’s not a novel claim. Where the thread split was over whether to treat it as an abstract threat or, as in reply 24, to separate “already-documented harms like discriminatory classification, fraud, and privacy violations” from “unverified scenarios like human extinction.” The former is real, present-day harm from current AI; the latter is a future prediction — and mixing the two tends to make the discussion talk past itself.
※This article is compiled and summarized from the 5ch (Business News+) thread “y‘¬•ñz‘˜AAAI„‚èul—Þ‘¶–S‚̃ŠƒXƒNvŒx [“c™ŽR–¬š].”
Leave a Reply