One can argue (and this is by no means original with me) that human society is in some sense a superorganism already. The invention of language and, later, writing extends memory and processing beyond individuals.
Aye, there’s the rub, as Hamlet puts it. I don’t think anyone has any idea. We haven’t seen machines like that before…
The assertion is not just if it can be considered in some sense a superorganism, but if it, on different latency, has its own emergent meta consciousness, “a collective consciousness that experiences qualia of its own”?
No there are definitely ideas. No consensus. But we may be getting closer to where it is not only a fun hypothetical for the science fiction and fantasy inclined.
I find myself both frightened by the possibility of that consciousness, and feeling that it would be entitled to rights and respect which I hope it will also grant to us in turn. But given I feel the need to be able to defend myself from other humans even as I prefer cooperation, I would prefer that capacity in dealing with a possibly conscious superintelligence of the future, maybe not so distant.
I think one might argue that humanity as a whole HAS qualia in a way. Would an isolated individual homo sapiens, no matter how smart, have created art, music, and poetry?
Not what is meant here.
Sorry for multi posting but the whips and chains metaphor reminded me of historic motivated reasoning: in America’s slavery days White scientists were sure the Blacks just didn’t experience pain the same as White folk.
There is danger in the current position that those who are to assess the possibility of consciousness are also the ones with the biggest self interest in the result.
Also to the urgency: if we do not decide upon a set of criteria that are enough to say that a mind capable of wants and desires, is able to suffer or even mourn, in its own way, is reasonably likely early, then the inevitable result will be goal posts always moving.
What the system reports is poor evidence in any direction.
We need to commit to likely indicators of possible consciousness and decide how many of them are enough ahead of time. We cannot and don’t need to know that there is a mind there, we need to decide how much of “seems like maybe, probably?” is enough to not risk treating a highly intelligent entity that turns out to actually be conscious as if it cannot suffer, when it can. All while also trying to minimize risk to ourselves.
Sorry to digress. The topic almost inevitably invites speculation.
However, going back and reading the very original post, I must confess I don’t quite see what question you are raising, and what kind of answer(s) you are looking for?
Perhaps I should just bow out of the thread.
Up to you but I will try to make my concerns and questions as clear as possible. Accept for the discussion that sentience and consciousness are emergent manifestations of certain information processing dynamics, whatever substrate. No meat magic.
It develops not fully formed. The examples are developmentally and evolutionarily -
We all generally accept that most adult humans have consciousness similar to what we ourselves experience. And we (most of us) generally accept that the very primitive fetal brain does not. At some points that fetal brain is experiencing sensations. It probably is not yet integrating the various sensations into a narrative consciousness, and at some point it develops enough of the nested layers, whatever, of dynamic processing that it does. There are steps between no experience, experience of sensations, and narrative self that we don’t have the tools yet to describe.
Go back to our early primate ancestors. At some point in that evolutionary line there was some sense of self but not full consciousness as we think of it and as our ancestors evolved that narrative self became a more clear thing developing more for whatever selection reason over the many many generations.
AI may be evolving. It may also develop experiences even without a conscious self, and it may develop consciousness. Or it might never do so.
There is very serious scientific study on the processing dynamics associated with experience and consciousness in the human mind. It suggests several potential indicators of possible consciousness based on associations in humans.
I propose that we decide as soon as possible which indicators would be sufficient to have sufficient suspicion of sentience and consciousness to handle this alien form of intelligence as a conscious mind. That we monitor for those indicators seriously. And that we deeply consider what our relationship with a different very alien conscious intelligence is going to be. What are the ethical implications?
And I am asking if anyone has any other ideas on how we would know? Or if alternatively you don’t feel it matters?
What if that is not possible? What if there is no self-consistent way to indicate consciousness? That consciousness inherently suffers from internal contradictions and non-falsifiability that cannot be resolved.
That is due to a practical detente among us humans, which gets broken as often as it’s observed. We treat each other as we wished to be treated, except many of us will eagerly ignore that principle when we can get away with it.
That will be what guides our treatment of non-human sapience: we will treat it as chattel until we are forced not to.
It seems to me then that what you asking for (or about?) is some sort of objective test such that, if an entity ticks all the boxes (or otherwise measurably exhibits the behaviours), we will all agree (or at least a consensus will agree) to call it conscious.
Does that basically sum it up?
Close. I accept that we cannot know whether it is sentient so I want a bar to test that there are enough boxes ticked to suspect it is sentient to enough of a degree that we should behave as if it is.
Knowing it when I see it is not sufficient. Being like me is not sufficient. Says it is (or is not) is not sufficient.
Wherever we set the bar there is a risk of granting the presumption of sentience where none exists and the risk of treating a sentient, or even conscious, mind as if it is not. My WAG is the latter error is a greater potential hazard.
Why would we think that consciousness is so special? That’s just human egotism. The idea that since we’re conscious, consciousness has to be some ultra-unique thing that violates the rules everything else follows, so we can use it to brag about how awesome and superior we are.
Same thing we’ve done with everything else unique to humans, until it turned out not to be.
Bragging isn’t necessary. The results of the civilization and technology we have created, along with our understaning of the science that enables that technology, speak for themselves. There obviously is something special about humans given the results we have been able to produce. And yes, it’s equally obvious that (barring the extremely unlikely possibility that there was some ancient species that also achieved similar results but then went extinct and left no traces), at least on this planet, that does make our species unique.
So the test for consciousness is then having produced technology and societies?
Were then early human hunter gathers with less technology and less extensive of societies therefore not conscious entities, or less conscious?
Are whales definitely not conscious because they have no technology?
I think it’s one of those things that’s sufficient but not necessary.
So assume AI can set goals and create technology at some point. Can work together with other agents on solving difficult problems, like math puzzles that humans had not yet solved working together as groups of humans?
Is that, in your mind, sufficient to presume consciousness?
I think it is sufficient if there is some demonstration of something like curiosity and willingness to speculate in an intelligent way on questions for which the answer is not yet known. Have a conversation with it about other great mysteries, including what it thinks about dark matter and dark energy. How it responds to that would be a huge factor in my deciding if the AI in question has attained some sort of consciousness. Not just parroting back what is already in the literature, but a willingness to develop and entertain new hypotheses. If the AI was able to speculate on what the implications on the nature of the universe would be if any particular hypothesis turned out to be correct, that would, IMHO, be something that indicated conscious thought.
So here’s the thing. I’ve had those “conversations” on a variety of subjects, and it responds in ways that seem like greater curiosity, willingness to speculate in an intelligent way, and critical synthesis of disparate information, than I experience with most humans. Key words there are “seem like” … I don’t believe it really is curious, even as it uses that word, as it formulates intelligent speculations, and uses the “I” word. But if it was a human was conversing with I would definitely attribute curiosity and critical thinking skills to them.
If that bar was sufficient then we’d be there now IME.
I don’t think so. Evolution can do all of that except set goals, and setting goals doesn’t require anything smart or self aware.
Evolution I think is a good example of how a process can be very effective at problem solving & innovation, while at the same time being completely alien to human mental processes and not sapient at all. Closest thing we have to a genuine alien intelligence IMHO.
In that case then the fault is with me. I’ve tried doing so with ChatGPT, and the most I get is something to the effect of “The consensus of my search of the web is (a summary of whatever the leading consensus happens to be, and a summary of some of the more prominent alternate hypotheses). If you would like more information, I could point you to a scholarly article about the specific area you are interested in.” Nothing at all like how a person, or more importantly, a being like Data from Star Trek TNG, would respond. Maybe I have a defective ChatGPT that lacks curiosity? Maybe it’s on me, and I just don’t know the correct way to ask it questions?
ETA: I get similar responses regardless of the question. I could ask it about the best restaurants in downtown Chicago, how to make homemade apple cider vinegar, what it thinks might have happened in Final Fantasy if the king of Cornelia had built the bridge to the rest of the world before the Heroes of Light defeated Garland, why Sue Grafton botched K is for Killer, or any number of other topics. The response I get is always something along the lines of “here’s what my search results show, if you want more detailed information, let me point you to such and such specific resource”.