We don't have enough information yet so we can't make an informed decision. It's possible but we are nowhere near. It's going to be at least another generation just to get started. But that doesn't mean that what we have will not have a great impact on society. It will. Starting yesterday!
n. the ability to derive information, learn from experience, adapt to the environment, understand, and correctly utilize thought and reason.
1. Fact. A computer, by definition, can only evaluate computable functions. For instance, it cannot decide the halting problem.
2. Hypothesis. The set of all functions that form the basis of intelligence is computable
3. Fact. It is not known that Hypothesis 2 is true.
4. Therefore, it is not knowable that a computer can be intelligent or never be intelligent unless Hypothesis 2 is proven.
The statement It is possible that a future computer could be intelligent is false because it assumes Hypothesis 2 is true could be proven true in the future. So, it is conjecture at best.
The statement Some computers are already intelligent is false since Hypothesis 2 is not proven to be true yet.
The statement A computer can never be intelligent by definition assumes that Hypothesis 2 can never be proven to be true. This is a very strong statement without basis. Therefore, it is also conjecture.
The statement Every computer is intelligent by definition is false since it treats Hypothesis 2 as an axiom. So, it can be true only in a universe where Hypothesis 2 is true.
We know the brain is doing information processing. Our senses are just streams of information mapped to our nerve signals. Our brain is constructing our environment, our memories, and our inner feelings. Damaging certain areas of the brain interrupts the information processing and predictably impairs specific functions associated with that area.
While we don't know the specific set of functions that form the basis of intelligence, we know there is information processing going on and therefore intelligence is computable. We don't know exactly the abstraction level at which neurons are forming this theoretical Turing Machine, however unless there are any reasons to think the brain is doing more than just information processing, we can conclude that machines will also be intelligent.
Then there is the definition of intelligence which I think lacks precision. There are many definitions but to continue with the provided definition, for example, "thought" and "reason" are not precise and is presuming a biological entity as the only ones having the capacity to be intelligent. Either we make that an explicit condition, meaning the definition of intelligence is something that biological brains do, or we define it such that it can apply to anything even non-living entities. If we do the latter, I think large GPT models are approaching a form of intelligence. The Turing Test is soon to be surpassed with large GPT models, where most people will not be able to tell the difference, and at that point we will have to admit that it's at least a certain type of intelligence, although perhaps not human.
Piccinini, Gualtiero & Scarantino, Andrea. (2010). Computation vs. Information Processing: Why Their Difference Matters to Cognitive Science. Studies In History and Philosophy of Science Part A. 41. 237-246. 10.1016/j.shpsa.2010.07.012.
leaves the question open on whether cognition is computation or information processing? while laying out an exposition on the difference between computation and information processing.
The last two sentences in their paper's conclusion is:
> Thus, cognition may well involve the processing of information. Whether this information is processed by means of (digital) computation remains an open question. (reference to footnote xii)
and in the footnotes (number xii):
> One of us believes the answer is likely to be negative (Piccinini 2007b).
If by intelligent you mean capable of subject experience, there is nothing in known physics that even vaguely suggests that it could lead to subjective experience.
Simply responding to events in meaningful ways is something entirely different. There doesn't seem to be any limit to that beyond what is practical.
8 comments
[ 3.2 ms ] story [ 28.6 ms ] threadn. the ability to derive information, learn from experience, adapt to the environment, understand, and correctly utilize thought and reason.
The statement It is possible that a future computer could be intelligent is false because it assumes Hypothesis 2 is true could be proven true in the future. So, it is conjecture at best.The statement Some computers are already intelligent is false since Hypothesis 2 is not proven to be true yet.
The statement A computer can never be intelligent by definition assumes that Hypothesis 2 can never be proven to be true. This is a very strong statement without basis. Therefore, it is also conjecture.
The statement Every computer is intelligent by definition is false since it treats Hypothesis 2 as an axiom. So, it can be true only in a universe where Hypothesis 2 is true.
Therefore, my response is None of the above.
[1]: https://dictionary.apa.org/intelligence
n. whatever inorganic computers can't do yet.
[1]: https://youtu.be/iOVbAmknKUk
While we don't know the specific set of functions that form the basis of intelligence, we know there is information processing going on and therefore intelligence is computable. We don't know exactly the abstraction level at which neurons are forming this theoretical Turing Machine, however unless there are any reasons to think the brain is doing more than just information processing, we can conclude that machines will also be intelligent.
Then there is the definition of intelligence which I think lacks precision. There are many definitions but to continue with the provided definition, for example, "thought" and "reason" are not precise and is presuming a biological entity as the only ones having the capacity to be intelligent. Either we make that an explicit condition, meaning the definition of intelligence is something that biological brains do, or we define it such that it can apply to anything even non-living entities. If we do the latter, I think large GPT models are approaching a form of intelligence. The Turing Test is soon to be surpassed with large GPT models, where most people will not be able to tell the difference, and at that point we will have to admit that it's at least a certain type of intelligence, although perhaps not human.
The last two sentences in their paper's conclusion is:
> Thus, cognition may well involve the processing of information. Whether this information is processed by means of (digital) computation remains an open question. (reference to footnote xii)
and in the footnotes (number xii):
> One of us believes the answer is likely to be negative (Piccinini 2007b).
[1]: https://www.researchgate.net/publication/222414469_Computati...
Simply responding to events in meaningful ways is something entirely different. There doesn't seem to be any limit to that beyond what is practical.
the meaning of intelligence is up to the poll responder and it's an implicit part of the poll