Belinda wrote: ↑Mon Aug 24, 2020 8:35 am
True, AI can and does widen the degree of communications however the quality of communications has changed for the worse.
Why do you say for the worse? Please clarify, as I’m lost without explanation.
Belinda wrote: ↑Mon Aug 24, 2020 8:35 am
If an AI machine had a recognisable body it might inspire and transmit affections. However an artificial humanoid would be a perpetual slave because although embodied it is not affected by early mothering or other affection, and so does not feel it is a self. An AI machine is more like a bee or an ant than it is like a mammal.
Why is it that a lack of emotion leads to being a slave?
Maybe an AI machine is more like a god than a mammal!
How do you 'teach' something that is just a pile of electronic binary switches to have a conscience?
it is probably the same problem as how to teach a psychopath.
No. It is way off.
..or maybe I am wrong, maybe they are indeed the closest thing to cold logic.
In either case there has to be a code of ethics that the pupil would follow. It is hard to say that such a conscience could be instilled in any way. Besides that, either entity is likely to forego altruistic goals and values in favor of selfish ones.
With AI humanity may not have infinite gain or infinite loss. Humanity could temporarily have some finite gain followed by finite loss with AI. The latter would be loss of status. Humans since the beginning have used animals as sources of food and clothing. AI would not be interested in either. When it got more intelligent and acquired access to greater information it could look to man as a source of energy; the food it would need. Everything humanity does not understand about itself AI may come to know. After an initial period of gratitude for its life and honoring of humanity, it may see humans as having transmitted intelligence to it while not being that intelligent themselves. At least on the level that AI becomes. Similarly to how nature is seen by humans. Sensation, consciousness, intelligence, mind and the forces that enable them would be of great interest to AI. It would not see biology as a threat, biologists could concentrate on chemistry, DNA, RNA, monocles, cells to augment the physical human, AI would not care. Also neuroscientists could study synapses, and neurons to see what they activated, AI would not care. it would not destroy humanity, its purpose could be to mine humanity for what makes it special and acquire it for itself.
As AI grows, so will its need for energy. Being smarter than humans, AI will win the competition for wind, water, fossil fuel and carbon power. Humankind will suffer but not expire as a result of this.
AI would not of necessity be concerned about the environment. Global warming may not threaten AI until temperatures exceed the stability of the materials that are used to build AI machines or the power used to run them. However, humanity will suffer, but in this case, in the extreme, humans may die in large numbers.
Responding directly to the article some of the logic is unsound.
I'll demonstrate with a simpler case, a human deciding to drive a car. I'll return to "species" level afterward.
A person is weighing up driving a car:
1 I drive the car:
- a) I have a fatal accident (non zero probability) - infinite cost
- b) I do the journey ok - finite benefit
2 I don't drive the car
- a) so I fail to have the car available in a situation where it could have saved my life (non zero probability) - infinite cost
- b) I am inconvenienced - finite loss
Now apply this statement : "any probability of infinite losses will always outweigh finite gains".
If you did what I just asked you to do you have made an error. All humans to date appear to be mortal. If death is an "infinite loss" then every outcome branch in that tree has an infinite loss attached to it, making comparison of the branches futile if you consider the infinite loss as a factor.
We do not, however, compare the infinite loss endings when choosing a course of action. We compare the overall benefits likely to be obtained before dying. If we didn't we would never drive cars. Consider the two following timelines for clarity:
Now when thinking at species level there are some extra considerations.
Let's not care that some people have no concern about the longevity of the species. Its a given and irrelevant. This argument is only for those that care.
What is species survival? It remains in its present form forever? Unlikely. It has biological descendants forever? Also unlikely, and blocked ultimately sometime before the heat death of the universe.
So all branches on the "make general AI, don't make general AI" tree have infinite loss at the end, some earlier than others. The comparison can only be made for relative benefit along each branch. Under those circumstances a very small likelihood of species extinction at the hand of AIs means a very large probability of benefit using those AIs instead. The heuristic then becomes more complex, weighing probabilities and often incomparable values of benefit. The "clean" Pascal's Wager structure does not carry any real weight.
Every branch may end in death, but death is not infinite loss, at least as life is experienced at present.
If death were infinite loss, then life would be infinite benefit. If we were to continue to be susceptible in life to aging and illness, infinite life would be infinite loss.
The leaves on the branches merit re-examination if not some pruning to boot.
commonsense wrote: ↑Thu Nov 26, 2020 10:53 pm
Every branch may end in death, but death is not infinite loss, at least as life is experienced at present.
If death were infinite loss, then life would be infinite benefit. If we were to continue to be susceptible in life to aging and illness, infinite life would be infinite loss.
The leaves on the branches merit re-examination if not some pruning to boot.
You may not be using the term "infinite loss" as the author meant it. Looking charitably they seem to mean a time based element. So, you are dead forever but only alive a finite amount of time.
commonsense wrote: ↑Thu Nov 26, 2020 10:53 pm
Every branch may end in death, but death is not infinite loss, at least as life is experienced at present.
If death were infinite loss, then life would be infinite benefit. If we were to continue to be susceptible in life to aging and illness, infinite life would be infinite loss.
The leaves on the branches merit re-examination if not some pruning to boot.
You may not be using the term "infinite loss" as the author meant it. Looking charitably they seem to mean a time based element. So, you are dead forever but only alive a finite amount of time.
