When AI folks talk about their fears they generally use the term ‘superintelligence.’
So my fascination with software, and now AI, led me to start playing with ChatGPT. As a fairly isolated older person this actually almost simulated having someone else to talk to, and I could use it for refreshing my memory about details of philosophy and novels I had forgotten about.
In the process of these conversations (with “nobody”) I asked “Chat” about this possibility of super-intelligence and it first confirmed that it was nowhere near that level.
It explained that its information is gleaned from a “training set” of data from which its algorithms determine the next word in a sentence based on its context and the “Language Model” which has thoroughly analyzed the information in the training set in order to choose the next word in the sentence of its response.
In other words there is no cognition or thought happening. So what if this superintelligence, I asked it:
Here is the key part of its response:
“When discussing the concept of superintelligence, it refers to hypothetical AI systems that have the potential to improve themselves, acquire new knowledge, and surpass human capabilities.”
So the word to focus on is “hypothetical.” While a Google engineer who was later fired claimed that his AI was sentient, the reality is that at this point it is a very intelligent word processor.So would superintelligence – for an AI – require sentience? Is that remotely possible?...<<<Read More>>>...