h2g2 The Hitchhiker's Guide to the Galaxy: Earth Edition

Florida Sailor,

Many imaginative and creative narratives have been offered to explain disease in different cultures in different times. The difference with the germ theory of disease is that careful appeal to the data yielded predictive and explanatory power.

So I wonder what it is that you are asking about AI in this scenario? Are you considering whether AI could make the creative or imaginative leaps that seem to characterize some scientific breakthroughs?

>>Are you considering whether AI could make the creative or imaginative leaps that seem to characterize some scientific breakthroughs?

I think that is exactly my point. If it is truly intelligent on a human level It should be able to break beyond the obvious information it is given. A student in school is taught many facts that they can repeat like a parrot. A true education teaches them how to research, experiment and create new facts by themselves. If a machine is truly intelligent it should be able to tell us more than it was told by us.

F S

Florida Sailor,

Thanks for clarifying.

I don't know whether human level AI is possible or if it is, how far into the future its first development will be.

I agree with you that creativity would be an important element in such a system and I think this raises interesting questions about the nature of human creativity.

First, thanks to Robbie for appreciating the conversation; I can’t imagine anything worse than having to give a talk about AI to Global Law Firm partners but, presumably, someone has to do it.

Florida Sailor’s point about creativity is an extremely important part of this (and my apologies for being somewhat flippant about it in post 56, real life has been rudely interrupting my train of thought for a few days). I think what you’re talking about is intuition. The human individual can only exist in a world, a framework made up of significant connections (yes, I know, I’m trying to avoid the word “matrix”) and, although we may not be altogether conscious of it, we are aware of that world of connections enough to be able to jump around in it to reach conclusions. Obviously, sometimes it will be the wrong conclusion but a lot of the time it’s correct. At present, no machine is capable of having that individual identity that entails having a world so they have to follow basic logical steps.

Here’s an interesting piece about this: http://www.bigthink.com/21st-century-spirituality/black-box-ai

Hi folks! OK ... creativity. There are various different kinds of creativity ... being able to prepare an appealing meal; designing a building; writing music or singing it; creating visual art; writing stories or poems; creating new scientific theories. A am fairly sure that creativity is indeed something some people are intrinsically better at than others ... which is not to disparage others; creativity is just *one* attribute of intelligence. Also I can't say if the greater creativity of some people is due to a difference at the level of the genes, or whether it 'emerges' at a higher level! (I'll get back to mind as an emergent phenomenon or epiphenomenon ... for now I'd just like to apologize since I did actually confuse the two.)

In my own case, it was evident that I was very creative when I was four or five years old; I just spontaneously started painting and sculpting and could do things I still marvel at today. And this artisticness of mine appeared fully-fledged, practically instantly. There has to have been some 'innateness' involved: I immediately turned out good-looking people and animals after virtually no practice work at all.

But there are other aspects of art I'm not so good at. Actually in my brain I'm far more creative than in anything I actually paint ... for the physical painting I still am struggling with hand-eye coordination as I manoeuvre my pencils and brushes, perceiving shapes, getting the exact correct colour.

So ... I get it all the time that people compliment me on my art talent. Sometimes it dismayed me, as if talent is all it takes, but in fact my talent alone wouldn't bring me far, I have to engage in very hard, regular, sustained work so as to keep producing and to keep getting better ... I'm still nowhere close to what I'm aspiring to.

And I think an AI could not be called intelligent unless it, too, had ambitions and aspirations, and if it, too, is able to dedicate itself to learning/growing/getting better.

Another point about creativity. There's easy creativity and hard creativity. The easy kind is simply coming up with novelty. That is easy ... one of the first things I did after we got our Commodore 64 comouter was that I started writing programs for creating sentences, by choosing from a selection of verbs, nouns, prepositions etc. Even that ancient computer could with this program create highly original, and often hilarious, sentences. This is 'creativity' but surely no actual mind is involved!

*Hard* creativity is creating something that is not merely novel, but also conforming to a higher standard: truth, beauty, harmony, inspiration. And teaching *that* to a computer or AI would be the big challenge!

Chris, thanks for that piece! There, they're speaking of AI's running important stuff and I agree we can't leave such important decisions to machines or computers whose 'thinking' we don't quite understand. I feel that this is separate from the sphere of scientifically trying to make a truly thinking AI. That falls under pure research ... while we might create monsters, at least we can pull their plug in time.

Willem,

I agree that there are different kinds of creativity that fit the standard definition, that to be creative there must be originality and effectiveness. For example many musicians have been creative within an idiom but relatively few have played a significant part in developing a new idiom.

If I interpret you correctly, your conviction that some people are more naturally (you have now said 'intrinsically') creative than others is based partly on your own experience. I think it is very difficult to draw such conclusions from our own memories of how we developed our own creativity though. The findings of cognitive psychology suggest that we don't necessarily have access to the many hours of experimentation and play that preceded what casual observers think has come from nowhere, even at a very young age. This matches my own experience as a guitar player, I know I spent many hours learning to change between chord shapes, but I have no direct memories of doing so.

Given the thread topic this might be worth a look, it gets off to a slow start but does get interesting:
https://www.youtube.com/watch?v=h0962biiZa4

Thanks for that link, I haven’t watched the whole thing yet but I was so annoyed by what Elon Musk was saying about half way through that I had to comment.

First I was surprised at what a primitive level a lot of the thinking was in general. Elon Musk has a view of human consciousness that was common when the social sciences still aspired to a reductionist model, that it proceeds through a simplistic, mechanical process and anyone not displaying such mechanical thinking is failing to live up to proper human standards. He also voiced his concern that if AGI was achieved by a single state or group it would give dictatorial power to the entity without, apparently, recognising that we already live in such a world and that he himself is already part of just such a power elite.

Anyway, I’m going to watch the rest of the video now so perhaps I’ll have changed my mind by the end.

ust updating the link of those of who still use ripley http://www.youtube.com/watch?v=h0962biiZa4

Chris,

Funnily enough I'm only half way through it myself. One interesting thing said was relevant to the black-box article that you linked to above, that AI can be developed to explain its own reasoning in a way that humans can understand.

It was an interesting article you linked to thanks, although one minor criticism I have is the idea that ethics "are relative to the culture and time they are being practiced in". This is a complicated area, but I would certainly reject some versions of moral relativism that could be inferred from this comment and to question whether we can program morals into AI on this basis seems to me to be a non sequitur.

I'd be interested to discuss your objections to what was said about mechanistic views, as I'm not sure I understand them.

First, I’ll post what I’ve managed to copy down from the subtitles of Elon Musk’s answer (starting about 32:30 into the video). I haven’t actually listened to it so I have to presume the subtitles are accurate.

“So, we have to figure out, what is a world like to be in where there is this digital superintelligence? I think another point that is really important to appreciate is that we are, all of us, already cyborgs. So you have a machine extension of yourself in the form of your phone and your computer and all your applications. You are already superhuman. By far you have more power, more capability, than the President of the United States had thirty years ago. If you have an internet link you have an article of wisdom, you can communicate to millions of people, you can communicate to the rest of Earth instantly. I mean, these are magical powers that didn’t exist not that long ago. So everyone is already superhuman, and a cyborg. The limitation is one of bandwidth. So we’re bandwidth-constrained, particularly on output. Our input is much better but our output is extremely slow. If you want to be generous you could say maybe it’s a few hundred bits per second, or a kilobit or something like that output. The way we output is like we have our little meat sticks that we move very slowly and push buttons, or tap a little screen. And that’s extremely slow compared to a computer which can communicate at the terabyte level. These are big orders of magnitude differences. Our input is much better because of vision, but even that could be enhanced significantly. So I think the two things that are needed for a future that we would look at and conclude is good, most likely, is we have to solve that bandwidth constraint with a direct neural interface. I think a high bandwidth interface to the cortex, so that we can have a digital tertiary layer that’s more fully symbiotic with the rest of us. We’ve got the cortex and the limbic system, which seem to work together pretty well – they’ve got good bandwidth, whereas the bandwidth to the additional tertiary layer is weak. So I think if we can solve that bandwidth issue and then AI can be widely available.

…if it was limited to a small number of people and it was ultra-smart, they would have dominion over the Earth. So I think it’s extremely important that it be widespread and that we solve the bandwidth issue. And if we do those things, then it will be tied to our consciousness, tied to our will, tied to the sum of human individual will, and everyone would have it so it would be sort of still a relatively even playing field…”

However, it will take a while to write a proper response to this and Psi’s post.

Not to beat a dead horse, but let me go back to my post 55 'The 19th Century View of Fever'. I think this might be a good example of something AI might be capable of. This was a serious problem that scientists struggled with for hundreds of years. We did not know that mosquitoes where responsible until Walter Reed made his finding in the early 20th century, then we learned the truth, fever is spread by mosquitoes . There were several clues in reports made about 60 years earlier. The quotes below are from early 19th century reports, the comments are my own.

>Fever was rarely seen when the temperature dropped below 40°F (4°C). As the temperature rises the frequency and severity of fever increases until it reaches a plateau at about 100°F (38°C). People living near deserts were almost exempt from most fevers.

The activity of mosquitoes are also reduced, or eliminated in cold weather, or a dry climate.

>Fever was most prevalent in marshy areas where there was an abundance of decaying vegetation. It was observed that fever was fairly rare in forested wilderness areas, but cases increased dramatically when trees were felled and fields ploughed.

These are also the areas that mosquitoes breed.

>A good example of the effect of elevation is Rome, Italy, where fever was common; in nearby Tivoli some 300 feet higher in elevation, cases of fever were rare. This same phenomenon could be seen in many areas of the world.

Mosquitoes feed on mammals and flightless birds such as chickens and therefore fly close to the ground looking for food.

>It was believed that the fever gases were carried by breezes and currents of air, not all obvious or understood. In some cases a home that appeared to lie in an exposed position would be free of fever, while another house far removed would be stricken severely.

It makes sense that a Mosquito, light and full of energy would fly into the wind, so that when she was tired and gorged with blood, she could return home with a tail wind.

>In her book Notes on Nursing published in 1860 Florence Nightingale opposed the common idea that sick rooms should be closed to fresh air:
The only time when it can be unsafe to open the window at night is when the air is more foul without than within. This may be the case in close back courts, and in malarial countries, or at hours when there is a sudden fall of temperature. But even in malarial districts it is found that thin gauze curtains, while admitting the air, are a protection from Malaria. Always air your room, then, from the outside air, if possible, Windows are made to open; doors are made to shut...

It sees obvious (at least to us) that the gauze screen would eliminate insects, including Mosquitoes.

It sees to me that a computer with a large enough data bank, and not confined by false conclusions might have been able to evaluate this problem before humans could. I am not suggesting it could have solved the problem, just showing us what the examine.

F S

Chris,

That transcript tallies with my memory of listening to it.

Psi

Florida Sailor, This would definitely qualify as one of the best uses of AI and I think someone, possibly Sam Harris(?) mentions it on the video, but as far as I can make out we already have the sort of AI that can handle this type of problem-solving. The problems kick in when we start thinking about AI that can think about us thinking about them.

Sorry it’s taken me a long time to write this but the questions Psi raises in post 70 are extremely difficult even to answer superficially so I’ll restrict myself to areas that I think are directly related to this conversation.

I’ll deal with my objections to Elon Musk’s answer in the Future of Life video first because, although the questions are related, this one is slightly more straightforward to answer.

First, I recognise that he is answering “off the top of his head” a question thrown at him so he should be allowed some leeway, but it is his area of expertise and it would seem odd if he had never thought about it before. However, if he hasn’t and we take his answer at face value, it does seem to support the view, explicit in post 64 but implicit throughout all my posts in this conversation, that individual human being means having a world. Entrepreneurs such as Musk and Richard Branson, for example, succeed by jumping around in their world, grabbing disparate ideas and throwing them together without analysing closely where they come from or what they mean. As a couple of examples of how I think this works, many years ago a friend doing an Eng Lit degree said she had an assignment to write a poem so I offered her a Limerick about her favourite poets, one line of which went “And Sylvia Plath was a gas, gas, gas” (come on, Limericks are meant to be in bad taste!). I vaguely knew that this was quoting something but it took a long time to remember that it was a Rolling Stones lyric, and Ian Anderson of Jethro Tull apparently took years to realise that the line in Aqualung, “Do not start away uneasy,” was paraphrasing Robert Burns. What I think this means, as far as the idea of mechanistic thinking and Musk’s view that our output is restricted, are concerned, is that, actually, we have no output; we are already in the world (meaning both our individual world and physical reality) and it’s only when we “artificially” attempt to step outside ourselves and analyse what we think or do that we produce any output (and this “output” is actually not output but something being put in to our world).

If we do take Elon Musk’s answer at its most basic level, then his assumption that everyone has a similar life to his is clearly wrong; there are an estimated 1.3 billion people with no access to electricity never mind the internet and probably more than that with no access to dependable food and water supplies. So, no, everyone is not already a superhuman cyborg and, further, his idea of technological augmentation also indicates a lack of awareness; my daughter was born with a condition that caused her to become deaf and, although she now has a bone-anchored hearing aid that is at the cutting edge of medical and electronic technology, these things cost almost as much as our house. The only way that she can have her hearing aid is through the NHS and no matter how superintelligent AI becomes, I don’t think it would ever be capable of inventing a national health service. More interestingly, why would it need to? We are already clever enough to think of it ourselves even if we don’t have the political will to fund it properly. On the other hand, if we assume that Elon Musk is at least aware of a more general philosophical argument as the subtext to his answer, then this becomes a much wider subject that embraces the questions Psi poses on the article by Derek Beres.

The philosophical view underpinning Musk’s idea of the autonomous individual struggling to connect with anything outside itself derives, mainly, from Descartes as I’ve suggested in posts 3 and 17 and, while this allows for precise modelling using basic propositional logic, it fails to address the problem of Cartesian Dualism. The initial attempts to derive some absolute truth that seems to be promised by the Cartesian project followed two different paths towards understanding how we come to have knowledge of physical reality; Rationalism which distrusts any evidence that we obtain with our senses and Empiricism which holds that knowledge can only be found outside ourselves. But both of these views are essentially materialist and so, in both there is a tendency for the subject to become impotent or disappear altogether. Therefore, the possibility of extending the objectivity of science to the mental and social sphere has so far been unsuccessful, and the range of proposed analyses and solutions to that failure are as numerous as the micro-categories of Heavy Metal music. However, most of them, in some fashion, follow from Hume and Hegel and, although Hume was an empiricist, he took it to its illogical conclusion (as Viv Stanshall used to say) and showed that there were profound problems in its logical foundations so that both of those sources tend to result in some form of relativism. As MacIntyre has pointed out, in his typically MacIntyre style, “Hegel and subsequent historicists were right in claiming that morality which is no particular society’s morality is to be found nowhere,” the point being that, as long as this remains open to change, not taken to be true always and everywhere, it is not inconsistent with relativism. Terry Eagleton comments that “It is a striking feature of modernity that we find ourselves unable to agree even on fundamentals. Almost everyone takes the view that attempting to asphyxiate small children is not a course of action to be recommended, but we cannot agree on why we agree on this, and perhaps never will.” (Terry Eagleton on Wittgenstein’s politics, Commonweal Magazine January 2017) and, according to Adorno, the point is not so much to reconcile subject and object in the sense of bringing them back together, but rather, to reconfigure them in a state of differentiation without domination. So, although none of us apart from a few weird French intellectuals, really want relativism to be true, it’s maddeningly difficult to find a convincing argument against it and, when Beres writes that the moral world of a North Korean is different from that of a North American, he is simply describing the situation as it stands at the moment.

John Donne told us that no man is an island, but I would suggest that having an individual identity is both the constant struggle to become an island and the necessity of failing to become such. Apart from anecdotal evidence of Eastern mystics achieving some form of independence from material reality but, presumably, losing their individual identity to some universal life force, to succeed in becoming an island is to die and there is a recognised medical condition, Cotard’s Syndrome, in which the sufferer becomes so detached from the world that they believe they have died or that they no longer have a self. Jules Cotard’s original patient thought that she had lost all the organs in her body and must be immortal, therefore she didn’t need to eat and starved to death. Because we have a body, we are not physically differentiated from the world; our body consists of more bacteria, viruses, funguses and mites than it does human cells and material reality is constantly passing through us to keep the process going. Since the 17th century we have two means available to us of attempting to describe the inside and outside of being human and the relationship between them – modern science and ordinary language. The former, based on the materialist duality of the 17th century and as Willem has noted, has struggled to see the reflexivity of the subject; the latter has ambiguity as both an advantage and a disadvantage. In researching this post I came across Donald Davidson, an American philosopher of whom I was not aware (there’s thousands of them – every time you lift a rock a dozen more pop out of the mud) and his idea of Anomalous Monism which gives a very convincing logical account of how every mental event can be paired with a physical event thus “SOLVING THE PROBLEM OF DUALISM!!!” This slogan has been around for a long time now and I sometimes feel that it’s a bit like McDonald’s, if you tell people their meal is Happy often enough it becomes The Real Thing. What people like Davidson seem to be doing is inserting words into a space that may or may not exist and, while trying to use language as precisely as possible, depending to some extent on the fuzziness of language to keep that space slightly out of focus. Of course, if the words do in fact create a space that didn’t exist before, then that would prove relativism to be absolutely true which would, naturally, prove relativism to be false.

You may have spotted by now that what I’m doing here is revealing my “political agenda”, the ideological view that informs my opinion which is that there is an unresolvable contradiction underlying what it is to be a self-aware individual because I want Free Will to be The Real Thing but, if science comes up with some irrefutable evidence that the mind is simply another cog in the machine then I have to hold my hands up and say “OK, I was wrong” which my wife informs me I am most of the time. And I have been reading about a small area of the brain, the Claustrum, which is “hidden away” underneath the insular cortex, made up of large cells with dendrites covered by spines that are capable of both receiving an input and projecting it back. Not much research has been done on it because it is so difficult to isolate but it is possible that it could act as a “Cartesian Theatre” or as an orchestral conductor keeping all the instruments in time to produce music rather than cacophony. Also, in 1964 Kornhuber and Dreecke discovered what they referred to as “Bereitschaftspotential”, that about 550 milliseconds before the subject had any conscious awareness that they were going to act, their brain was already preparing to initiate movement. These findings were confirmed by Libet’s experiments in the 1980s.

Meanwhile, I’ll cling on to my contradiction because, for me, it explains the importance of music and humour; they are two of the ways we have of juggling contradiction, keeping those contradictory balls in the air at all times. As that great philosopher Loudon Wainwright III has written, in a post-modern world “Time and Life and People are just glossy magazines.”

Chris,

Superb post, I had to scurry off to my copy of After Virtue to read the quote in context. I think there is a way to reconcile this with the direction the panel including Musk took the discussion.

A few years ago someone labeled my stance as 'Davidsonian' so I did read some, but again I will need to scurry and check 'Anomalous Monism'.

It will take some time and care for a detailed response to your thought provoking post though, so please bear with me.

After Virtue - the book that introduced me to the idea of post-modernism. It's all downhill after that!

I always have to think, 'What would Slavoj Zizek say about that? And how many sniffs would it take to say it?'

I like that concept of Bereitschaftspotential. I shall have to look this up.

Hi folks! I'll try to get back to this ... would like to explain my own position re minds, artificial intelligence, etc. For now ... I do think we're going to see artificial intelligence comparable to human intelligence. Perhaps not soon. But we might have centuries and millennia still in which to do it - which will, geologically speaking, not be long at all. We need to solve our pressing present problems first though. But the benefit of creating artificial intelligences is that on the way there, we're going to learn about our own minds much more than we know now. So it will be worth it and will also have lots of side benefits. Anyways here's how I envisage a 'real' artificially intelligent entity ... something that starts out much like a human baby but has the potential to learn over time and develop personality and character. Not something that we make to solve practical problems for us - as 'artificially intelligent programs' right now, or programs to trick their way through the Turing Test, but as electronically-based true minds - to be considered as being *people*, therefore. Such an artificial mind, and artificial *person*, will have to be given rights and respect for being what it is. Such an artificial mental being will have desires, goals, ideals - and I'm hoping the potential to become a 'good person' ... or perhaps a bad one! The whole idea of potential is that it can have potential for good or for evil. But if evil ... we'll have to figure out did we make it evil or did it become evil of its own accord - and if so, do we punish it? How? Anyways there are all sorts of interesting conclusions we can make about mind by *pondering* the possibilities of artificial minds. That's the interesting aspect to it for *me*.

Willem, I'm interested in hearing your opinion of how an artificial intelligence would develop aspirations and desires without being part of a living body, with relatives, community and history etc.