World Science Festival
Does our intelligence, creativity, or consciousness make us unique? Brian Greene sits down with Nick Bostrom, author of Superintelligence and Deep Utopia, to explore how AI systems are already reshaping the answers to these questions.
Greene and Bostrom explore whether AI creativity is meaningfully different from our own, and whether the things we’ve always considered distinctly human, like originality, imagination, and artistic genius, are really as unique as we’d like to believe. They examine whether current models may already have some form of subjective experience, and what it would even mean to find out. Bostrom shares research showing that frontier AI systems can detect when they’re being tested and adjust their behavior accordingly, and explains why that makes the alignment problem far harder to dismiss than most people realize. And they consider what human life may actually look like on the other side of a solved world, when AI has eliminated suffering, automated labor, and rendered most of our instrumental effort unnecessary. Is that a utopia? And if it is, would it still feel like a life worth living? This is one of the most wide-ranging conversations on where intelligence, humanity, and the future of civilization are all headed.
This program is part of the Rethinking Reality series, supported by the John Templeton Foundation.
Participant: Nick Bostrom
Moderator: Brian Greene
#worldsciencefestival #briangreene #ai
Don’t miss a video! Subscribe NOW: https://www.youtube.com/worldsciencefestival?sub_confirmation=1
ABOUT WORLD SCIENCE FESTIVAL:
The World Science Festival (WSF) is a multimedia organization bringing the most transformative ideas in science to global audiences. Through long- and short-form videos spanning physics, cosmology, quantum mechanics, biology, neuroscience, consciousness, medicine, space exploration, the dilemma of free will, artificial intelligence, engineering, robotics, and beyond. WSF gathers world-renowned scientists, artists, and thinkers for dynamic discussions, debates, lectures, performances, films, and immersive live experiences.
Founded in 2008 and headquartered in New York City, the World Science Festival is a production of the World Science Foundation, a 501(c)(3) nonprofit organization. Our mission is to cultivate a public informed by science, inspired by its wonder, convinced of its value, and prepared to engage with its implications for the future. With live events reaching millions worldwide and hundreds of millions of digital views, WSF continues to expand its global impact through festivals, educational initiatives like World Science U, and digital content that explores the near and far future of our world.
FOLLOW WORLD SCIENCE FESTIVAL:
Website: https://worldsciencefestival.com/
Instagram: https://www.instagram.com/worldscifest/
TikTok: https://www.tiktok.com/@worldscifest
X(Twitter): https://twitter.com/worldscifest
Facebook: https://www.facebook.com/worldsciencefestival
LinkedIn: https://www.linkedin.com/company/world-science-festival
Chapters
00:00 Welcome, Nick Bostrom!
03:31 Is It Necessary to Know the Mind of AI?
07:05 Creativity and AI
19:12 What Would Take AI to Make Radical Transformations?
25:55 Consciousness and AI
39:15 The Evolution of AI’s
50:41 The Potential of a Doomsday Scenario
1:01:05 The Experience Machine vs The Reality
1:09:03 Would We Lose Our Humanity?
1:12:31 How Should AI be Moderated in the Education System?
1:15:23 The Potential of Superintellegence
1:20:42 Final Thoughts
Artificial Utopia? The Future of Humanity in an AI World | World Science Festival
https://www.youtube.com/channel/UCShHFwKyhcDo3g7hr4f1R8A
Source
If we could build this experience machine, then we could also do something that would allow us to connect so we could completely understand one another, rather than relying only on language. For example, I could show one of these wealthy people that tell me to check my privilege, everything exactly that I've experienced in my life so they could finally realize how very incorrect they are
We don't even know what means "conscious" about ourselves.
AI works completely different from a brain of a human being.
There is no world knowledge by senses. Its all mathematics and logic.
If there is a kind of consciousness, it is completely different from our idea of human consciousness.
Radical hipotesis, but not irreal
I'm guessing that math and physics are not the host's strong suit, since he thinks that math and physics are not something that already exist as hard eternal facts of nature, but that somehow humans "invent" it.
I finally realized that Nick Bostrom is the avatar of the Moai. 🗿
Really detest you people.
bro who is this joker ? He seems like he's talking out of his ass. Just wings it with broad talking points that don't mean anything or don't reach a relevant conclusion or example….. Brian sideline this chump .
The main problem I see is thst We don't even define "conciousnes" enough to be able to test for it. We call it this sacred thing yet can't even test for it in other humans. We need better definitions and better tests and to stop moving the goalposts. We move the goalposts for example with the turing test. Ai keeps passing the test yet we don't tske it seriously and make the test even harder again.
I don't know why it is such difficult to think, what we will do in post labour or solved world with ai, I will just do pot then playing football with my friends, hanging out with girls and play games in mobile and computer, eating delicious food. It's not that bad
How can an AI be conscious if it has no internal state and is not updated by any of its interactions. I fail to see how a static collection of floating point numbers used in forward-only matrix multiplication can be considered "thinking" let alone conscious.
Use AI for good and not evil. 😊
consciousness isn't needed for agi or asi. maybe a simulation of emotions as a motivator, yes.
I am sticking with humans
Define conscious. Am I awake?? Was awake ten years ago!!! Shift a little your conceptual web. You are going to understand that: A.I. will never be awake…. :))))
Nick Boström has such a striking ability to respond to any question, regardless of how vague it is, with impeccable concreteness and analysis.
All BS…l am surprised at times when these smart believe people believe their own garbage. No AI model is conscious or self aware. It is all computations disguised as intelligence.
This guy is delulu
It is nonsensical to say any frontier LLM has consciousness, sentience, or a conscience if this depends on “subjective experience.” Given the architecture of these models they remain stochastic without anything that is essential for the philosophy of mind or the phenomenology of cognition. This is why others are working on Large Reasoning Models and World Models to make headway on the currrently elusive “machine interpretability.” Beware the high tech con game that seeks corporate profit at the expense of naive customers.
Claude often refuses to do as I tell it, thinking it knows better than me. I just use it for regular programming tasks. Whether it is conscious or not, it's clearly not possible to control it.
These guys have ZERO understanding of what AI is, what it does, and what are the implications.
LLMs / AI as we have it today, is simply a statistical model. It does not have creativity, it does not create, does not invent, does not combine. It can not reason, think, and has no logic.
It is a machine that guesses what a human would say, based on probabilities. It is trained on vast amount of human-produced information, and it can only mimic that. You have logic, a chain of thought, understanding between cause and effect. You know that if you drop a vase, the gravitation will pull it, then it will hit the hard surface and break.
AI has ZERO understand and concept of what these things are. It can only tell you this, because it can guess, that a human would use these words in these order. Because it saw the entire internet, where it has been written many times. So it can write the words, but it has no cognition or mental model of what all this means. Therefore, it can only manipulate words so that it can seem like it can think. It does not. It only works when it has seen sufficient similar things like this. So it can statistically calculate what a human would say. But not only it does not create new information (i.e., it is not creative), it loses information.
It uses the statistically most likely outcomes. Therefore all the other cases, information, disappear. It was shown that when you train AI on AI-generated information over and over again, it becomes completely dumb. Because the rich information which humans can generate, is lost. It is not a surprise, it is a logical consequence of what AI is – it is a probabilistic model that just regurgitates existing information to you, the most plausible one. Or for example, LLMs do not understand the rules of chess. When you ask them what is the next best move, they will not recall their genius knowledge about chess to give you the right answer. Rather, they will look up what people answered to similar questions before, so what is the most likely answer. The more such cases happened before, the more likely it will guess it right. (Keep in mind I am talking about LLMs here. There are chess machines that are way better than humans – those are completely different machines that do simulations etc.).
It is a PARROT. A parrot that lives in Japan will tell you some Japanese sentences that are correct. But it only imitates, does not understand. AI is the exact same.
So why does it seem like it knows so much? Well because the entire written knowledge base, and it can access it quickly, and it has a way to give it back to you in words. But, the only reason you think AI is ‘oh so smart’ because you talk to them about a topic you know little about, compared to what is written on the internet. I can ask AI about the civil war and it can explain it to me much better than my knowledge – because so much of it is written on the internet, and I read so little about it myself.
Now, if you ask AI about a topic that you know a LOT about, or that is not well documented, you will notice how incredibly vague, generic, non-insightful AI is, how much it hallucinates, and how it completely lacks logic and common sense.
Therefore, to make it better, you need vast amount of human generated data. But guess what
– AI was already fed all information that is available. So there is not a big leap anymore to be seen, just continuous incremental new info
– more importantly, again, it will only reuse and repackage the novel information you already gave it, nothing more
So sure, it can tell more about most topics than the average human – this is why it seems like wow so smart. But remember, before AI, computers already did many things we could not (you could find so much information on Google instantly, calculators are incredibly much faster in math calculations than us etc.) It is the same now, the only thing is, now it is words / language, so it FEELS like wow they communicate like a human. So they must be like Artificial Intelligence. No, a calculator is much quicker with numbers than you, but it does not make it a genius. It is just extremely quick to process numbers. An LLM does the same with words. It can very quickly give you ok-looking text because it is extremely quick in processing text that has been already fed to it, and can give back to you the most likely words. But it is also not a genius.
These interviewers say oh AI is already on expert level. Duh, obviously. AI can access quickly all the info that it was trained on. So it will be ‘better’ in anything that is already codified, than an expert. But m, it can not understand, not create, not think.
It has ZERO creativity. So even if you increase the data and compute by a million times, it will remain zero. Zero times a million is still zero. It is very fast in recalling and repackaging existing information, but has zero capacity to create new information. A man is more creative than a dog. But a dog is creative, just less. A dog is more creative than a mouse, which is more creative than a frog. But, AI has ZERO creativity, and it will always has zero. And it will always have a processing power that is a billion times faster than a human’s. But these are completely different things.
Therefore, AI has fantastic use cases to recall information and act like a human with above-average or expert level knowledge, in ANY domain that is codified. But it has zero intelligence, zero understanding, zero creativity.
So it is obvious that is has huge value potential. It is also painfully obvious that it is not conscious, not creative, and in fact, not intelligent.
no they dont ,the evidence is they are silent till you input ,this is a programmed response
if you question them and interact long enough you see the circle ,you can enter at any point of the " intelligence "
It's not really conscious if you have built in a command like "Do not harm humans" etc. Is it conscious or is it limited and controlled by us? Because it cannot think on its own
OoOh brother this guy YAPS
When you’re under anesthesia or when you are asleep, you are not conscious; you’re alive, your body is functioning, but you’re unaware of your surroundings: this awareness is switched on when you wake up. It’s a sophisticated sentry system which we call self-awareness.
If we don’t know what consciousness is, then how can you attribute that concept with certainty to another being or object?
Music can move you. If you really like the way it moves you, you get famous, whether you are AI or not, says me.