Do LLMs Understand? AI Pioneer Yann LeCun Spars with DeepMind’s Adam Brown.

Janna Levin (Professor of Physics and Astronomy) 00:00.620

the is um when when they really are self-motivated agents if that ever actually happens, that they won't collude, fight amongst themselves, want to wrestle for power, that we won't be sitting back watching conflicts that we simply couldn't have imagined

Yann LeCun (Chief AI Scientist) 00:19.340

before. We give them clear objectives and we build them in such a way that the only thing they can do is fulfill those objectives. Now, this is not doesn't mean it's going to be perfect, but the question of AI safety in the future, I'm I'm worried about it in the the same way

Yann LeCun (Chief AI Scientist) 00:33.900

that I'm worried about. The question of reliability of turbojets,

Yann LeCun (Chief AI Scientist) 00:38.340

okay? I mean turbojets I mean it's it's amazing. I don't know about you, but and my dad was aeronautical engineer, but I'm totally amazed by the fact that you can fly halfway around the world in complete safety on a two-engine airplane. It's amazing, right? And

Yann LeCun (Chief AI Scientist) 00:55.440

and we feel completely safe doing this. It's a It's It's a magical production of engineering of the modern science and engineering of the modern world. AI safety is a problem of this type. It's It's an engineering problem.

Yann LeCun (Chief AI Scientist) 01:10.400

I think the fears are caused by people who think about you know, science fiction scenario where somewhere someone invents the secret to super intelligence, turns on the machine and the next second it takes over the world. That is complete BS. Like the world doesn't work this

Yann LeCun (Chief AI Scientist) 01:28.880

way. Certainly the world and technology and science doesn't work world this

Yann LeCun (Chief AI Scientist) 01:32.160

way. The emergence of super intelligence is not going to be an event. Um as we see we have super intelligent systems that can do super intelligent tasks, you know, and there is kind of continuous progress when at a time. Uh,

Yann LeCun (Chief AI Scientist) 01:48.400

but you know, we're going to find some, you know, better recipe to build AI systems that may have kind of a more general intelligence than we currently have and and we'll have system. There's no question that are smarter than humans. But we'll build them so that they fulfill the

Yann LeCun (Chief AI Scientist) 02:04.800

goals we give them, subject to God rails.

Janna Levin (Professor of Physics and Astronomy) 02:08.400

Um, I I I was going to uh again question this idea of we we we we know that if we can code them in a certain way, somebody could re-code them. And the concept of bad actors, but before we fall into that hole, I have a plant in the audience. Does my plant have a mic?

Janna Levin (Professor of Physics and Astronomy) 02:28.080

Is my plant know who he is? Does my Meredith, Isaac, does my plant have a mic? Yes? Oh, but he doesn't have the mic. Okay, David, can you shout?

Janna Levin (Professor of Physics and Astronomy) 02:41.480

Okay, so um so I want to introduce the uh philosopher of mine, David Chalmers. I'm going to give you a very brief introduction. David, I can't see you but I I I said um that you could be my plant to ask a question. Could you Do you want to throw something down

David Chalmers (Professor of Philosophy and Neural Science) 03:01.000

here? Okay, I didn't know here. Okay, you Jana asked me to ask a question about uh AI, consciousness, higher atom. Hi you Okay, so uh you both said I think roughly current AI systems probably not conscious

David Chalmers (Professor of Philosophy and Neural Science) 03:21.120

future AI systems possibly descendants of the one today but some future AI systems probably will be conscious. So I guess I'm going to know that one what requirements for consciousness do you think current systems are lacking

David Chalmers (Professor of Philosophy and Neural Science) 03:38.640

um and then the positive side of that is what steps do you think we need to take in order to Well, AI systems which are conscious and then third when is that going to happen?

Yann LeCun (Chief AI Scientist) 03:55.000

Okay, I give a quick at this. And David already knows my answer but so first of all I don't attribute like I don't really know how to define consciousness and I don't attribute much importance to it.

Yann LeCun (Chief AI Scientist) 04:09.360

And this is an insult to David, I'm sorry. Because he devoted his entire career to it.

Yann LeCun (Chief AI Scientist) 04:17.280

Okay, that's a different thing. Okay, subjective experience. So clearly we are going to have systems that have subjective experience that have emotions. Emotions to some extent are an anticipation of outcome.

Yann LeCun (Chief AI Scientist) 04:31.560

If we have systems that have world models that are capable of anticipating the outcome of a situation, perhaps resulting from their actions, they're going to have emotions because they can predict whether something is going to end up, you know, good or bad for you know, in on

Yann LeCun (Chief AI Scientist) 04:48.120

the way to fulfilling the objectives, right?

Yann LeCun (Chief AI Scientist) 04:50.080

So so they're going to have all of those characteristics. Now, I don't know how to define consciousness in this kind of in in this, but perhaps uh consciousness would be the ability for the system to kind of observe itself and configure itself to solve a particular sub problem

Yann LeCun (Chief AI Scientist) 05:06.200

that it's facing.