Weekly Poll: Will (Should) We Create Artificial Intelligence (A.I.)?
A popular subject of research today is in the nature of consciousness and whether computers can be made to replicate human thought. The 'holy grail' of this research is the creation of A.I., Artificial Intelligence — computers that can think independently in ways that go beyond the initial programming or foundations provided by human creators. If humans manage to do this, it will help us understand so much more about the nature of consciousness and thinking than we currently know. Will this succeed? Should we want it to succeed?
I'm sure we've all seen at least some of the many stories about the possible negative consequences of creating Artificial Intelligence, but those stories are usually more allegories about current situations than genuine warnings about likely risks, aren't they? It's implausible that the creation of A.I. is 100% risk-free — every scientific and technological advance carries risks of some sort, even if sometimes minor. Every time we learn how to do something, we open the door to people doing something improper with that knowledge. We shouldn't get too carried away, though, with dystopic visions of a future laboring under robot overlords.
On the other hand, there are a host of ethical issues that could be unleashed by the development of Artificial Intelligence. Would intelligence make a machine sentient? Would an Artificial Intelligence have rights or duties? Could it be allowed to vote? What if it wanted to reproduce and create new Artificial Intelligences that would now lack direct human input or controls? Eventually A.I. could become cheap enough to install anywhere, but would you really want your toaster or television to be invested with a sentient intelligence? What if it refused to let you watch programs that are insulting to intelligence?
Religious reactions to the development of Artificial Intelligence would be interesting. On the one hand some would surely use it as "evidence" that an intelligence requires an intelligent designer; on the other hand, this would definitely be evidence that an intelligence doesn't require anything supernatural, transcendent, or divine as an explanation. Believers in souls would be faced with the question of whether an Artificial Intelligence has a soul. If not, it would be harder to defend the belief that humans must; if so, it would be harder to defend the idea that souls are supernatural and immaterial.
I wonder what an Artificial Intelligence would think about such debates?
Comments
Some have said that this has already been done. It’s just not human artificial intelligence.
If by human intelligence we are talking about the machinations of the actual human brain, I would not think A.I. could actually duplicate the thought processes of a sentient human mind. Our brains are infinitely complex organisms. Science hasn’t been able to completely map the mental processes even now, with our scientific knowledge.
The human brain is still a mystery for the most part. Perhaps, in a time many years from now, when in a distant future, the brain’s mysteries have been completely solved, true human-like A.I. may happen. For the forseeable futeure, I would think not.
For instance, could we install a chip of some sort for love, sadness or happiness? Emotions are uniquely human characteristics of our thought processes. Are those possible to replicate? Would a synthetic emotion be an emotion?
I guess only time will tell if that’s even a remote possibility. For now, I am of the opinion that true A.I. is in the realm of sci-fi.
It is only a matter of time before human level AI will emerge. It doesn’t necessarily follow that all machines would then possess “personalities” any more than we would call our thermostats or microwaves “computers” even though they technically are.
Intelligence will be distributed throughout networks and in most cases the “sentience” would emerge gradually and imperceptibly. It’s possible that the web itself already has a form of sentience.
To the extent that we would allow durable self-aware personalities (as opposed to transient programs) to exist, they would have rights. Certain applications would be advantageous, such as when you wanted a machine to provide expert opinions based on its experience.
But sentience is a two-edged sword. Once a machine begins to have opinions, it is no longer fully under human control. We would also not want such a sentient device doing the drudgery of repetitive computation. So like a thermostat, we will continue to employ a vast majority of computational AI in a non-sentient manner, choosing carefully when we want to create a machine intelligence which might talk back or disagree.
As for your discussion of the implications of AI for the concept of the “soul.” Bingo. Once it’s incontrovertibly established that consciousness (i.e. all human thoughts and feelings) are mechanistic processes it will forever dispel notions of mind-body dualism. Some people will remain unconvinced, but I’d say within 50 years they will be as universally ridiculed as “flat-earthers.”
Do I believe a thinking being (like a machine equipped with artificial intelligence) can be, at any time, be produced from inorganic mass? No.
Humanity evolves along side the technology. The possibility of a computer smaller than a room was preposterous until humanity evolved a need for it. Most everyone lived and died inside a 30 miles radius for eons until we invented the transporation means not to.Now almost no one does. AI presents the same stimuli to our species. It is altogether probable that future generations will find life as inconcievable without robotic interaction as we would without telephonic communications or flush toilets today.
AI might be the next step in our evolution. Unless we can find a way around the speed of light, we as a species are pretty much screwed. We’ll ruin and over populate this planet soon. We don’t have the means to terraform our neighboring planets before that happens. We don’t have any back up planets, and if we found one, we couldn’t reach it in time.
That is of course if SkyNet doesn’t get tired of us.
i, for one, welcome our new robotic overlords.
To me, the possibility of creating genuinely intelligent machines is not in question. Depending on your definition of intelligence, it may already have been done. Sentient intelligence too is possible, but I don’t know how probable within the immediate future.
I voted on the basis of the very general question posed here that it worried me, and yet it is not the technology, of itself, that causes concern, but the exact nature of it and purpose to which it might ultimately be put. This is true of all technology. We have already seen many examples of elegant technology employed in unworthy and disasterous ways. The cavalier or purposefully harmful use of artificial intelligence could be incredibly harmfull.
To make something as intelligent as we are, we first have to do one of two things. We would either have to have a computer which could evolve its instruction set such that computer could change how it thinks or we would have to simulate the physics of an entire human brain (which IF we could do this, have we made computers intelligent or simply confirmed that the human brain is intelligent?). Searle’s Chinese room thought experiment is a powerful example of how computers (at least instruction specific computers; the only ones we have right now) cannot be made intelligent and any intelligence is simulated at best. AI is more the study of making computers SEEM intelligent, not making them think on their own. We may be able to write software to pass the Turing test or something that can seem very human in its interaction, but computers which can reason, create, express, and understand their world are not likely to be.
Knowledge marches on and we have very little to say about it. Darwin suppressed his discovery of evolution because of the controversy it would create but eventually published because others were about the publish the same ideas. Who actually invented the telephone? Who invented the airplane? Many people were working on these technologies at the same time. We can slow or alter the development of knowledge such as when we suppress funding for stem cell research, but people are still working on stem cell technology, the pursuit of knowledge continues.
Intelligence superior to our own will develop. Do we really think evolution ends with us? There are no transitional stages, every generation is the next stage, whether in biology or technology.
My bet is we will be superseded by a technological intelligence because technology evolves so much faster than biology.
I must disagree with you on one point good Mr. Sojourner:
“…[L]ove, sadness or happiness? Emotions are uniquely human characteristics of our thought processes…”
Ever had a dog?
While I cannot prove it, (nor most anything else discussed here), I’ve seen plenty of canines exhibit those very traits.
Most primates show basic emotions too (I recently read an article in Discover magazine about lemurs being visibly upset when given a smaller share of food than their peers – could that be rudimentary ethics?).