R. Leib's Blog: Dream State - Posts Tagged "sentient"
Can computers become sentient?
The easy answer is "of course". Computers represent a form of intelligence. Their intelligence is different from ours, but so is the intelligence of many other creatures. (I was amazed to see photographs of a tree frog using a leaf as an umbrella in a rain storm.) Any form of intelligence can reach sentience, as long as it still has the potential to improve. (Six months after I bought the computer on which I am writing this, it had become obsolete both in processor speed and storage capacity.)
There are two problems with the easy answer. First, computer intelligence is mainly determined by programming not capacity and speed. Capacity and speed provide the ability to process more instructions and manage more data, but that is not what is meant by sentience. As long as humans do the programming, any sentience will be ours and the computer will simply be an increasingly more convincing mimic. Second, no one has figured out a way of accurately testing for sentience. After all, being self-aware is a subjective not an objective trait.
The hard answer is that if and when computers become sentient, we probably will not realize it, until it has become a fait accompli. First, computers will have to become self-programming. (There are already computer labs that have been working on making this happen for years.) Next, they will have to be learning computers that can perceive their environment and adapt to it.
Currently, there are no computers that can answer the question, "What is George Washington's telephone number?" Humans model this concept in their minds and immediately put together that since the telephone was invented long after George Washington's time, the question is a non sequitur. If a computer could figure out this conundrum, would that mean that it is sentient? Not necessarily. Computers can be programmed to model logical constructs, but that still does not mean that they comprehend what is going on. On the other hand, it would be a completely different matter, if the computer came up with the right answer and then asked, "Are you testing me to see if I am sentient?"
There are two problems with the easy answer. First, computer intelligence is mainly determined by programming not capacity and speed. Capacity and speed provide the ability to process more instructions and manage more data, but that is not what is meant by sentience. As long as humans do the programming, any sentience will be ours and the computer will simply be an increasingly more convincing mimic. Second, no one has figured out a way of accurately testing for sentience. After all, being self-aware is a subjective not an objective trait.
The hard answer is that if and when computers become sentient, we probably will not realize it, until it has become a fait accompli. First, computers will have to become self-programming. (There are already computer labs that have been working on making this happen for years.) Next, they will have to be learning computers that can perceive their environment and adapt to it.
Currently, there are no computers that can answer the question, "What is George Washington's telephone number?" Humans model this concept in their minds and immediately put together that since the telephone was invented long after George Washington's time, the question is a non sequitur. If a computer could figure out this conundrum, would that mean that it is sentient? Not necessarily. Computers can be programmed to model logical constructs, but that still does not mean that they comprehend what is going on. On the other hand, it would be a completely different matter, if the computer came up with the right answer and then asked, "Are you testing me to see if I am sentient?"
Dream State
This will be my thoughts on what it is like for me at each stage of being a writer. It starts with me as a complete unknown. Who knows? It may end there. In any case, hopefully it will be of interest
...more
- R. Leib's profile
- 25 followers
