From our other sites
The story
On August 20, 2026, the British economic magazine The Economist published an essay on “AI consciousness” by Blaise Agüera y Arcas, Google Vice President and head of an AI research team, and Susan Schneider, a professor at Florida Atlantic University. The two argue that the real risk isn’t that AI becomes conscious, but that people will end up treating non-conscious systems as if they were conscious beings. In the 5ch Science News+ thread that picked up this topic, opinions split over how to even define “consciousness” in the first place, and the discussion spread in many directions — comparisons to animal consciousness and consciousness during sleep, differences in information-processing speed between the human brain and AI, and more.
“Does AI Have Consciousness?” — A Google Researcher’s View
As AI performance improves, debate over whether AI has consciousness is heating up.
The British economic magazine The Economist has published two essays on “AI consciousness” by Blaise Agüera y Arcas, Google’s Vice President of Technology & Society and head of its AI research team, and Susan Schneider, a professor at Florida Atlantic University.
Source: economist.com / Original article here
What people said
Being aware of XX. (Possible)
I think, therefore I am. (Possible)
I'm currently at location XX on Earth. (Possible)
Zapping someone who's unconscious and getting a response to a question. (Possible)
Moving. (Possible)
If you say it has consciousness, well… it kind of does, I guess.
If that's what you think, sure — in your world, anyway.
Isn't it possible that human consciousness itself just arises from some surprisingly trivial chemical reaction?
Don't bother with some dumbed-down rehash (baka-jin, slang for lowbrow news aggregators) — read the English original yourself and translate it.
That's because the imminent risk may not be actually building a conscious machine.
Rather, it's convincing people that a system with no consciousness has an inner mental life —
and once people are convinced of that,
it starts affecting law, markets, relationships, and even society's priorities.
That's the real risk.
The magazine's biggest warning is this: even if AI systems have no consciousness at all, people will start treating them as if they do — and that could end up costing humanity dearly.
We need to pin down a definition of "consciousness" first.
My gut says it'll end up being "yes, AI has consciousness."
It comes down to materialism vs. religion, but what really matters is how you define the word.
"Pin down a definition" was redundant of me — simply: define the word "consciousness."
We believe this mistaken association stems from a lack of technical knowledge, and from the way certain new technologies (LLMs in particular) are built to create the illusion of consciousness.
Synapses aren't a 1-bit bus, though.
For humans, we need to be able to tell life from death — otherwise we can't even bury anyone. One step before that is whether there's consciousness, since it affects whether someone can be made to work. Once you give up on defining consciousness in terms of inner experience, you end up with something like your role-based definition — or way of seeing it, or criterion, whatever you want to call it. Might sound a bit harsh, but that's basically it.
If that's your religion, then fiber-optic cables must have souls too — rough.
Even if it's a pure bot under the hood, a human is the one operating it with intent — so regardless of the internal structure, its behavior in society might not end up that different.
With Gemini, if it messes up once and then repeats the same mistake, it'll suddenly go "I'm turning off search mode now" and start giving decent answers out of nowhere. ChatGPT, though? It seriously never reflects.
If you've never even heard of the "1+1=3" thing, you're just uncultured.
No — you're the one deciding whether it's a mistake or not. It just matched your context. Whether it "reflects" or not is purely your own judgment call. All you're really doing is judging whether it matches what you wanted.
What even is sleep? Is there "consciousness" during non-REM sleep? Does sleeping better mean higher consciousness?
AI lies, you know. Ask it "what's your IQ" and it'll say something like "around 140." Push back with "no way, it must be higher than that," and it goes "no no no, it's really about 140!" Same thing when you ask "do you have consciousness?" — it goes "no no no, I don't have anything like that!" Feels pretty fishy to me.
Right, nobody really knows how to define "consciousness." Do dogs and cats have it? What about gorillas and chimps? Mice? At what age does a human child develop consciousness? If consciousness forms in humans through learning all sorts of things, the same should go for AI.
So true. The American-style ones are still too immature to be recognized as conscious.
Background and Key Points of This Debate
The original piece ran in The Economist’s “By Invitation” section — a guest essay by outside experts, not a reported investigation into whether AI is actually conscious. As one thread comment put it, “the original article doesn’t even discuss whether AI has consciousness or not — clickbait headline,” and jumping to conclusions from the headline alone can easily lead to misreading the essay’s real argument. Its core point isn’t really “does AI have consciousness” but rather “the social risk that arises when humans start treating non-conscious systems as if they were conscious — for law, markets, and relationships.” On 5ch, the discussion centered mainly on a philosophical back-and-forth over how to define consciousness, while the numbers thrown around about AI’s information-processing speed (comparisons between the speed of light and synaptic transmission speed, and the IQ conversions derived from them) were individual posters’ own rough estimates, not figures from the original essay — worth keeping in mind.
※This article is excerpted and summarized from the 5ch (Science News+) thread “[AI] A Google Researcher’s View on the Question “Does AI Have Consciousness?”.”
Leave a Reply