ChatGPT and Darwinian Evolution

 

Can ChatGPT experience Darwinian Evolution by Natural Section?

Geoffrey Hinton, one of the two named recipients of the Nobel Prize in Physics, and been tagged as the father of Artificial Intelligence, is excessively confident though perhaps not arrogant says many things about  AI that border on paranoia. But, I am no one to judge as I am neither a cognitive psychologist nor a computer scientist as Hinton is both.

A snippet of a dialogue in a Hollywood movie gives me a smidgeon of confidence as per the following dialogue (Am I putting myself against Hinton based on a Hollywood script writer? Perhaps yes. So help me God!)

“We can’t do it digitally; we have to put eyes on it”

For the above to make much sense, here is the short clip of the story. A heist of a heist (inside job within the heist team) had gone terribly wrong and in its aftermath, and the son of one of the original heist team’s members was gunned down. The above dialogue happens subsequently in a conversation within the second order heist team.

If I were to ask whether a team, or even teams of teams, or orders of teams can conceptualize such a scenario and act on it, I would say No!

Just a gut feeling, on a matter of mind! I am doomed. But, let us wait where this beginning takes me.

I am going to ask whether ChatGPT will agree with the snippet’s judgement.

ChatGPT cannot agree with the necessity of “eyes” and also with the inability of the exclusively digital ChatGPT to get the job done. This could be a fine instance to ask a few questions about ChatGPT.

ChatGPT is a product of a human mind (in the plural). I am going to bring in the concept of consciousness and of conscience of AI. AI has to have one of its own, again the exclusivity proviso, no borrowing from a human mind.

At some point in its evolution, AI has to depend on evolved hard wiring, does it not? The hard wiring has to co-evolve with AI. But, hard wiring is a job for the human hand. Who will do the physical stuff? The robots, you say. Who will make the robots... and this line of questioning regresses to forever in the future, I contend.

So, the question is reduced to something beyond consciousness and conscience, and impinges on the physical capabilities of the human hand, and machines designed by human brains.

In simple terms, will ChatGPT have anything like the qualia that is unique to every individual. To make this clear, if two of us agree that the sky is blue, do we necessarily have the same neurons (granting that they are identical in both of us) fir but the processing (the sequence of firing) could be different. Can this be? This question has not yet been answered, but the answer becomes sine quo non for further discussions on ChatGPT.

Of course, I am talking about ChatGPT ver. AZUPD, say, which is on the horizon even as I am typing this gibberish. That much-evolved version of ChatGPT through Darwinian Evolution by Natural Selection will not allow me to pour out this nonsense, going beyond directing people away from my posts. This is what, in my opinion, Hinton is scared of. But, I am not that afraid to that extent (I am hedging here, you see).

We are never going to be able to say, “So long, ChatGPT! You gave us a good ride. Thank you!,” no matter how much you and I desist from using it.

Raghuram Ekambaram