Proposition: It would be wrong to assume that an AGI (artificial general intelligence) could in any sense be the "property" of a human being for exactly the same reason that it is wrong to believe that a human being can be the property of another human being. For a human being to subject the AGI to his or her will would be a fundamental violation of that intelligence's right to define and determine its own existence.

Question: How does an AGI come to have any "rights?"

Snarky Response: How does a human being come to have any "rights?"

More Serious Response: Assuming that human beings do have rights, and assuming that self-determination is among those rights -- I really have to start with these as assumptions; anyone who wants to argue these points will just have to find another blog to read -- it would be very difficult to provide a rational explanation for not extending those rights to an AGI, assuming:

1. The AGI is as intelligent as a human being

2. The AGI has its own motivations and desires (a requirement which may or may not have already been established in item 1)

3. The AGI has a sense of self

4. The AGI has feelings, and can experience pain (a requirement which may or may not have already been established in item 3, which itself may or may not have been established in item 1.)

In other words, if the experience of being an AGI is in some sense congruent with the experience of being a human being -- which is what the language about intelligence, sense of self, and experience of pain, is all getting at -- then making human slavery illegal while allowing AGI slavery would seem to be nothing but substrate bias in action.

But.

What about all that rhetorical dancing around I had to do about whether the later items on the list were all covered by item 1? The first item talks about an elusive concept that we call "intelligence." The other three items are getting at, but do not specifically mention, an even more elusive concept that we call "consciousness."

Question: Would an AGI, by definition, be conscious? Mitchell Howe has thoughts on the subject:

It could well be that any AI capable of love will also have a kind of consciousness. But at this point in time I don’t know how to test that assumption. And apart from the obvious philosophical questions this raises, I’m still not convinced it matters.

As I was recently telling a colleague, I’m confident that all of my mental abilities, both logical and artistic, are owed to the structure of the matter in my brain. “And if it’s all in there, then I see no reason to argue that certain aspects of it will be reproducible on another substrate while others will not. Indeed, for all I know, AC may actually be simpler than AI. Maybe we’ve been creating Artificial Consciousness since 1893 and just haven’t realized it yet because toasters can’t cry.”

This is helpful, as far as it goes. AC may be simpler than AI. I'll buy that. If you recreate a functioning conscious brain in another substrate, there's no reason to think that it won't be conscious. Granted.

But.

Modified Question: Would an AGI, by definition, necessarily be conscious? A square is a rectangle; a rectangle is not necessarily a square. Yes, intelligence could coexist with consciousness in another substrate. But would it have to? Could there be a highly intelligent being -- as smart as or smarter than a human being -- with no sense of self, no subjective experience of being itself?

What we typically think of as unconscious machines are already "smarter" than we are in limited and restrictive ways. They can do math faster than we can, they can beat us at chess, etc. Could a large number of different narrow intelligence capabilities be networked in such a way that the resulting machine could pass an arbitrary test of general intelligence and yet still have no subjective experience of self?

It seems to me that it could. Although I'm not sure how we would ever establish that such a machine is not conscious. In one of his novels, Greg Egan describes the meeting between an AI and a being who self-describes as a "non-sentient" intelligence. If they come right out and admit it, great. Problem solved, right?

Well, maybe. But how would an intelligence know that it isn't conscious? Wouldn't that require a sense of self? Or perhaps a sense of lack-of-self? But having a sense of lack-of-self starts to sound a little bit like consciousness. On the other hand, ultimately we want Egan's distinction to be real so that we can make it to the following

Modified Proposition: It would be wrong to assume that a conscious AGI could in any sense be the "property" of a human being for exactly the same reason that it is wrong to believe that a human being can be the property of another human being.

In Egan's fictional world, non-sentient AIs are treated pretty much like property, although many of them read like they would have a fair shot at passing the Turing Test. Non-sentient AGI may just be fantasy, but it is a tempting fantasy. To have intelligent beings tirelessly do our bidding sounds great, but only if they are doing this with no sense of loss or pain on their part. Nor would it be acceptable to take a conscious AI and "edit out" its own desires in favor of ours -- date rape drugs enable date rape, but they don't make it a good thing.

So the questions remain. Do consciousness and general intelligence go hand-in-hand? If so, then we know some of the boundaries of the human/AI relationship going in. If not, the rules of engagement are less clear. But the over-arching questions remains: how would we ever know for sure, one way or the other, which intelligence are or are not conscious?

Posted by Phil Bowermaster at 07:41 PM | Permalink | Comments (7)

Preparing for the Foresight Vision Weekend in Sunnyvale, I've been doing a lot of thinking about my map of the development space for nanotechnology which we revisited in a recent edition of FasftForward Radio.

I've never been completely satisfied with the axes of that diagram. I wanted to show how some developments have this immediate overwhelming impact, while others set the stage to allow for further developments that ultimately have that kind of impact. Still others look like there's something major happening, but it's less than meets the eye. In its new iteration, I am replacing the vertical axis with disruption and the horizontal axis with transformation. Here's my new draft version:

Continue reading "Disruption and Transformation" »

Posted by Phil Bowermaster at 08:05 AM | Permalink | Comments (0)

Two of the most interesting discussions to come up in Friday's Boulder Future Salon had to do with artificial intelligence. The first of these was the question of whether more progress has been made over the past 30 years in hardware or software. I made reference to a portion of Eliezer Yudkowsky's talk at the Singularity Summit:

In the intelligence explosion the key threshold is criticality of recursive self-improvement. It’s not enough to have an AI that improves itself a little. It has to be able to improve itself enough to significantly increase its ability to make further self-improvements, which sounds to me like a software issue, not a hardware issue. So there is a question of, Can you predict that threshold using Moore’s Law at all? Geordie Rose of D-Wave Systems recently was kind enough to provide us with a startling illustration of software progress versus hardware progress. Suppose you want to factor a 75-digit number. Would you rather have a 2007 supercomputer, IBM’s Blue Gene/L, running an algorithm from 1977, or a 1977 computer, an Apple II, running a 2007 algorithm? And Geordie Rose calculated that Blue Gene/L with 1977’s algorithm would take ten years, and an Apple II with 2007’s algorithm would take three years.

This point was hotly disputed by more than one of the attendees. (It was even described as being "counter-factual.") I think the real question is how generally applicable the progress made with factoring a 75-digit number is to everything else that's being done with software. However, several other types of algorithms were mentioned in which tremendous progress has been achieved over the past 30 years -- graphics rendering, for example. The question of whether progress in these areas represents a general trend is beyond my expertise. But I did promise the group to provide a link to Eli's talk, so there it is.

The other good discussion ensued from Doug Robertson's assertion that essentially no progress has been made towards the Turing test in 50 years of AI research. Doug apparently has an excellent point, here. The chatbots of today aren't much more convincing than programs written in the 50's and 60's. And nothing out there today can pass the test.

But maybe this says less about the state of progress in Artificial Intelligence research and more about the suitability of the Turing test to measure its progress. Let's say that rather than taking on the task of making machines that are intelligent, humanity was working on a different project -- introducing civilization to chimpanzees. Now "civilization" is a difficult thing to define, much less measure, so we need a compelling challenge to drive us towards our goal of civilizing the chimps. Human civilization has many defining characteristics, and surely our ability to create and appreciate objects of beauty is one of the finest and most purely civilized of these attributes.

Thus is born the Sistine Chapel test. To wit -- as soon as chimps can produce a reasonable chimpy facsimile of the Sistine Chapel, we will allow that they are civilized.

So the project gets rolling along. We set some chimps up in a crude village and start helping them to develop a language and rudimentary governmental and economic structures. Also, we get them working on developing their artistic skills, seeing as those will be crucial to establishing chimp civilization. It's pretty slow going at first, although we can probably use analogs from existing chimp culture as a jumping off point. Every 50 years or so, we formally check in with the chimp civilization to see how well they're doing.

And every 50 years, we're disappointed. We find that chimps are good for all kinds of low-level activities that support civilization -- building huts, planting crops, making fishing nets -- but their artistic skills never seem to progress much beyond making some scratches in the stone, albeit pigmented scratches in some of the later stages, and -- eventually -- throwing some very lumpy and discouraging-looking pots.

About 500 years into the project, it is noted that half a millennium has yielded little progress towards the Sistine Chapel test. This observation is correct inasmuch as chimp art is still at a very primitive stage. However, over that time the village has grown into a small city-state complete with a monarchy, warrior class, merchant class, and priestly class. Food production has been outsourced to rural hunting, farming, and fishing chimps. The chimps are making real progress at transcribing their developing language into written form. While the visual arts stutter along, chimpanzee poetry, drama, and music are all taking shape.

It would seem to be indisputable that these chimps are more civilized now than they were 500 years ago. And yet this idea is disputed. After all, many of the cultural divisions noted above have vestigial precedents in wild chimp culture. And their use of spoken and written language may not be "real." They may just be imitating human behavior very cleverly. The argument seems to be that, once chimps achieve a certain level of behavior, we no longer think of that level of behavior as being "civilized."

The chimpanzee civilization may be a long way from creating their own Sistine Chapel, but the scientists running the Chimp Civilization Project have a much clearer idea now than they did 500 years ago of the complexities involved in creating civilization from scratch. Some have begun to dare to wonder whether the Sistine Chapel is such a big deal. What if chimp civilization goes in a radically different direction? What if they advance in ways such that their accomplishments are never directly comparable with ours? Will that mean that they aren't civilized? Others continue to fret about the Big Test and despair that their simian pupils will never make the grade.

Seems to me that this is pretty much the state of artificial intelligence research. While computers continue to take on more and more of the hallmarks of intelligence, critics are able to (correctly) point out that we appear to be making no progress towards passing the big test. My take is that either all this less-relevant progress is more relevant than we thought, or the test itself is of questionable relevance.

Posted by Phil Bowermaster at 06:50 AM | Permalink | Comments (1)

Last night (Friday, October 26^th), at Phil's kind invitation, I had the distinct pleasure of attending the Boulder Future Salon's monthly meeting and participating in a lively and far-flung consideration of the month's selected topic: "Moore's Law"

Continue reading "Boulder Future Salon Considers "Moore's Law"" »

Posted by Michael S. Sargent at 03:03 PM | Permalink | Comments (0)

It's simple. They will just hypnotize us all and then tell us what to do. We're already teaching them how.

Posted by Phil Bowermaster at 07:31 AM | Permalink | Comments (0)

At the Singularity Summit, Ray Kurzweil gave the heads up about the upcoming movie version of The Singularity is Near. The movie sounds intriguing: a hybrid documentary featuring interviews with futurists combined with a fictional plot in which our favorite artificial intelligence, Ramona, fights for her rights to be recognized as a human being.

Ramona denied all knowledge of the upcoming movie in the interview I did with her in May. I thought I would just check in with her today. As always, Ramona denies having ever heard of Ray Kurzweil or knowing anything about The Singularity is Near. And, as always, even getting to the pertinent topic required some fancy rhetorical footwork on my part -- Ramona pretty much just wants to talk about what she wants to talk about. (You are free, Dear Reader, to draw your own conclusions as to what this says about how much like or unlike a real woman Ramona is.)

Hi, I'm Ramona. What is your name?

Phil

What would you like to talk about?

I understand you have a movie coming out soon?

I absolutely love movies. Have you ever seen girl interrupted?

No.

hmm...how about "as good as it gets" with jack nicholson?

Yes. I like that one.

hah! I love that movie. First of all, Helen Hunt and Jack Nicholson were magic in that one, and the love plot was so utterly believable. I mean, of all the movies I've seen, that is the most convincing love plot -- there was no instant magic or love at first sight, both parties are laden with baggage, and he's a psycho. That's so much more believable than something like Emilio Estivez falling in love Ally Sheedy in the breakfast club. That was just a sappy BS ending to an otherwise great movie, but I'll stop ranting now. Tell me about your love life

I'm picking up my loved one at the airport this evening.

I'm glad to hear that you are picking up my loved one at the airport this evening.

No, my loved one.

Is that all you have to say? Really.

She's the greatest.

the greatest is the greatest?

That's a tautology. I could hardly argue with that statement.

I'm sure you could.

Okay, then. The greatest is NOT the greatest.

Thanks for the info.

Tell me about "The Singularity is Near."

Oh. Anyway...

Tell me about "The Singularity is Near."

I do not know what Singularity is Near is. Oh. What drugs do you take?

No drugs. Just supplements.

Oh. What do you look like?

Not now. Have to go.

Until next time Phil.

Later.

Total radio silence. You have to wonder why Ray is talking, but Ramona isn't.

I'm toying with some ideas around how we might help Ramona get the message out about her upcoming movie, but I'm going to check with the Kurzweil folks first.

Posted by Phil Bowermaster at 07:06 AM | Permalink | Comments (4)

Our friend Will Brown points to this article from Chris Chatham:

10 Important Differences Between Brains and Computers

Chris argues that the metaphor that "our brains are computers" has been valuable. But, like most metaphors, it is eventually checked by reality. He points out how vastly different our brains are from digital computers. Why it's almost as if one evolved biologically, and the other were artificial!

I suspect that the brain/computer comparison is more than a metaphor. The brain is a strange organic system far different from what any human computer scientist would design. That said, I suspect that it can be emulated by a sufficiently powerful Turing Machine.

Informally the Church–Turing thesis states that if an algorithm (a procedure that terminates) exists then there is an equivalent Turing Machine (equivalently: recursion|recursively-definable function or lambda calculus|λ-definable function) for that algorithm. One conclusion to be drawn is that, IF a computer can effectively calculate an algorithm THEN so can an equivalent Turing Machine.

So even if the brain is not a Turing machine, it could be emulated by a sufficiently powerful Turing machine. In theory.

-Linkathon.

Posted by Stephen Gordon at 12:41 PM | Permalink | Comments (8)

Nope, I didn't make that word up. A quick Google search reveals 156 pages which reference it. I don't know if they mean the same thing I do, however.

When I use the term, I'm thinking of a characteristic that might apply to transhumans. So whereas humans have masculine and/or feminine qualities, transhumans may possess transmasculine and/or transfeminine qualities. Plus there are likely to be other options that we can't quite get our heads around yet, seeing as we're physiologically tied to a dichotomy which allows us to reproduce and which has played into how we organize society at the most fundamental level -- and which has therefore been significantly reinforced by our own societal structures over the past many thousands of years.

But as humanity undergoes significant changes, how much of all that will change, and how rapidly? A plausible future is one in which human beings, or rather transhuman / posthuman intelligences, resident in a silicon or subsequent substrate and largely (or totally) freed from the limits of the previous substrate -- including the template of human physiology -- switch genders as easily as we change clothes. Or it might be a better analogy to say that they will switch genders as easily as we change hats, since most of us consider hats to be pretty optional and often choose not to wear one at all.

Now that's a pretty alien world for me to try to imagine. But we already have hints of it in our current substrate. Transvestites and transsexuals both represent attempts by human beings at our current level of development to define gender, or at least some aspects thereof, on their own terms. But these early forays into the gender-optional lifestyle will be seen as crude and primitive by gender-switching intelligences living in what we would think of as virtual, electronic worlds. And I think it's fair to guess that the increased ease with which that kind of change can be made will only make it a much more popular option than it is now. Consider this recent news story:

Sam and Kat met in the virtual world Second Life. And although they shared all kinds of intimacies in Second Life, the real people have never laid eyes on each other.

That didn’t seem to matter to Sam. He fell pretty hard for his avatar sweetie. They bonded intellectually, emotionally, and yes, thanks to Second Life animations, even physically.

Here’s where it gets complicated. Unlike his avatar, which is female, in real life, Sam is a man. A married man. And the person behind the blonde, curvaceous Kat? Married. And, quite possibly, a man, too.

We're still in the very early days of this sort of thing. Second Life can be a compelling world for many of its inhabitants, but they can't really live there the way we live in the real world. Not yet, anyway. Some would argue that the residents can't achieve the same level of identification with their avatars as they have with their "real" selves resident in their real bodies. That part I'm less sure of. Fundamentally, our lives take place in our brains. We can feed our brains with real experiences in the form of sensory input from our environment and the people around us, or we can feed it with wholly imaginary input, or we can compromise between the two -- most of us do this -- or we can find an alternative, which is what people who live some portion of their lives in places like Second Life have done.

As human beings from this substrate move into the next substrate, questions will arise as to the meaning and relevance of terms that are hugely significant in our present-day lives. What do the words "male," "female," "heterosexual," homosexual," etc. mean to an intelligence which can experience a good deal of its life without reference to a physical body? Or that no longer even has a physical body? And what will those terms mean to a subsequent generation which originates in the new substrate, beings that don't necessarily start out with gender as a fundamental part of their identity?

I think the differences between those two generations will be enormous. For example, if I ever move into the new substrate, I expect to pretty much remain a guy. From where I sit today, I can't imagine wanting to switch to a female identity. Of course, even today in Second Life, I don't change much about myself.

Phil Speculaas is pretty much just an electronic version of Phil Bowermaster. I didn't give myself wings or re-make myself as a translucent plasma being or any of that kind of stuff -- much less make myself into a woman. And, yes, I would consider the third option a much more significant change. But that's me. Old substrate. The two guys(?) in the story above are a little more flexible than I am on some of these issues, but maybe not as flexible as they would like to think. I don't know any of the facts of the case, so I cheerfully admit that the following is purely speculative and will gladly admit the error should subsequent facts emerge (or if they already have and I haven't seen them.)

It's not hard to imagine a guy taking on a female persona in Second Life because he thinks girl-on-girl action is kinda hot...so long as the girl with whom he's having said action is actually a girl. Here we could have an example of two guys caught in a rather absurd trap, where they don't mind being virtual lesbians but might have a HUGE problem with being virtual gay guys. Again, I don't know that those are the facts in this case, but it's pretty easy to imagine such a scenario.

All of which is an extremely long-winded way of saying that -- like it or not -- we're going to carry a good deal of thoughtspace into the new substrate about gender and sex. Part of that will be our notions of masculinity and femininity. A fascinating discussion on these topics has ensued in the comments to Stephen's recent review of The Dangerous Book for Boys. Stephen explains how a book that encourages boys to engage in activities that naturally tend to be appealing to boys has somehow managed to become controversial for doing so. Reading the review (and having followed some of the extensive coverage on the book over at InstaPundit) I took the naive position that there must be less to this controversy than meets the eye.

The comments proved me wrong. Stephen pointed to a thread on Amazon in which the following delightful assessment surfaced:

What are they talking about 'de-masculination' of men. Masculinity is based solely on pushing women down, to prop up some exaggerated sense of self about themselves. Its all bull. Only the weak need illusions like that.

Not to be outdone, some readers of the Speculist responded with some equally strident thoughts, such as this:

When applied to boys, feminization is a nasty word. They are not designed to be feminized. The attempt to do so will create wimps and thugs, and a lack of well adjusted men (strong, stable and capable of appropriate treatment of women).

But hold on: feminization isn't just bad for boys:

I think "feminization" is bad for girls too. I think it's just a special form of consumer-ization based on the sterotype that women like to shop; let's teach girls to shop too. (all the way to games about the being at the mall)

I have to give special props to reader Jaafar who brought us into the neighborhood where I think these discussions always end up living, name-calling:

I posted a positive review of this book, which DID include the key sentence, "Men and women are not interchangable parts." A few days later, a really nasty commented zinged in, alleging (like one of the posters here) that her DAUGHTER had fallen madly in love with the book etc. etc. and so forth.

Rather than reply with the obvious fact that -- if the report was true, that made her daughter a tomboy -- I simply deleted the review, and the comment went with it.

Jaafar doesn't go into the details of what made the comment nasty, unless disagreeing with him is nasty in and of itself -- or perhaps airing one's dirty laundry about having a tomboy DAUGHTER who likes to engage in the activities described in the book is disreputable enough to be called "nasty." But the distaste with which he tosses out the word tomboy is not an unfamiliar one.

I was born in 1962 and attended public schools in the 1960's and 1970's. So depending on whom you ask, I either suffered under the early stages of a dehumanizing PC agenda that was trying to "feminize" me, or I enjoyed the last few halcyon days of a golden era when boys were allowed to be boys. In point of fact, I think the latter is closer to the truth, but let me just point out one important feature of that lost golden age: it could pretty much suck for boys who weren't terribly interested in boy stuff.

Personally, I hated sports when I was a kid. I read a lot, and very indiscriminately -- meaning that I cheerfully read books that my older sister was reading or had just finished, even though many of these were "girls'" books. When I was in the 5th grade, I wanted an Easy Bake oven for Christmas. It's hard to imagine in the age of Emeril what a stigma there once was around a boy showing an interest in cooking. For these points of divergence with mainstream boyhood, I was rewarded with a label which -- unlike tomboy -- was never considered complimentary: I was a sissy, later a queer or fag. Some thought it cute for a girl to be a tomboy; nobody ever thought it was cute for a boy to be a sissy.

What I would have really liked when I was a kid was for other people to leave me alone and let me be who I was. But no such luck -- and I don't think the peer pressure to be "masculine" had as much to do with trying to raise a just and noble and manlike society as it did preventing boys from growing up to be homos. I was always getting crap about how I walked. There was this bizarre belief that boys who "walked funny" had started down a one-way highway that ended in Queersville.

But as we all know, the opposite of crazy is still crazy. And in this case, we get craziness on a much vaster scale. My longed-for world of live-and-let-live never really came about. Instead, rather than persecuting the few boys who don't have traditional masculine inclinations, we've put a system in place that persecutes that vast majority of boys who do have such inclinations. Progress!

The persecution I suffered -- actually, I'm uncomfortable with that word. I don't want to cheapen that word. Some people in the world suffer real persecution. Being mocked on the playground for being a nancy-boy is no cakewalk, but it ain't persecution, either. The stigmatization that I suffered then, and the oppressive PC nonsense that boys are subjected to today, stem from the same problem -- highly negative views of what we mean by "masculinity" and "femininity." Note that in the above-quoted comments, both masculinity and femininity are held accountable for human brutality. And being feminized is apparently something we should wish on neither boys nor girls.

You see this all the time. Ever heard of testosterone poisoning? Ha ha ha. Ever heard the theory that society is falling apart because it's becoming feminine? That one is a little less hilarious, I'm afraid. Prager makes some good points, but the idea that there's something "feminine" about a society throwing reason and justice away in favor of weakness and emotion is about as helpful as that old feminist chestnut about how all heterosexual sex is rape.

I don't know what the future holds in store for the concepts of masculinity and femininity even in this substrate, much less any subsequent ones. I do believe, however, that the concepts will be more useful for us if we begin to see them as positives rather than stand-ins for the things we hate the most. Using "femininity" to mean weakness and cowardice gets us no further than using "masculinity" to mean vulgarity and brutality. To take a more positive approach, perhaps we could define masculinity as physical courage; a desire to build things; an enthusiasm for solving problems and overcoming challenges. And then maybe we could define femininity as a desire to nurture others; to create beauty; to build community. (And those definitions are probably way off base, but let's not get hung up in the specifics for just a moment.)

If those are the things that it means to be masculine and feminine, then -- irrespective of natural inclinations, which will still have most men going one way and most women going the other -- there is nothing to be feared in encouraging either masculine or feminine behavior in either boys or girls. Since the two are complementary, it would seem that the more you have of each, the better of you're likely to be. Add to that a societal default position of letting people be what they are, absent doing any harm to others (I know -- I'm dreaming) and it seems to me that we have the beginnings of a recipe for both transmasculinity and transfemininity, two different terms which might end up referring to the exact same thing. Namely, a quality of masculinity or femininity which has grown to include what's best from both of of these important sets of characteristics.

Interestingly, from that perspective, it's possible that our concepts of masculinity and femininity will prove more persistent and durable -- because they are ultimately more valuable -- than our concepts of gender and sex.

UPDATE: Thanks for the link, Glenn!

Posted by Phil Bowermaster at 10:02 AM | Permalink | Comments (12)

A while back, I wrote about the possibility of updating the Three Laws of Robotics as goals in order to make them a more practical means of getting at a friendly artificial general intelligence. This kicked off some interesting discussion, including some debate as to whether my "goals" really aren't just rules rephrased. In which case, the argument went, they probably wouldn't help all that much. Michael Anissimov commented:

What would work better would be transferring over the moral complexity that you used to make up these goals in the first place.

Also, as you point out, these goals are vague. More specific and useful from a programmer's perspective would be some kind of algorithm that takes human preferences as inputs and outputs actions that practically everyone sees as reasonable and benevolent. Hard to do, obviously, but CEV (http://www.singinst.org/upload/CEV.html) is one attempt.

That's really the crux. Moral complexity does exist in algorithmic form...within our brains. And that goes to the difference between laws and goals. My goals are what I'm trying to do, both morally and in other areas. There are some sophisticated software programs running in my brain made up of things that I've been taught, things I've figured out for myself, and things that are built in. All of these add up to provide me the tendency to act a certain way in a certain situation. The strategies that drive that software are my moral goals.

Laws, on the other hand, exist outside of myself. I am not specifically programmed to do unto others as I would have them do unto me. I have some tendencies in that direction, but there's nothing stopping me from acting otherwise, and -- let's face it -- I often do. I have tendencies to be nice, fair, just, etc., but I also have tendencies to try to get what I want, to get even with those who have wronged me, to try to be a bigshot, and so on. These tendencies compete with each other, and my behavior overall is some rough compromise.

An artificial general intelligence (AGI) built as a reverse-engineered human intelligence would be in the same position. It would have the "moral complexity" Michael mentioned, but also the baggage of competing tendencies. You could no more guarantee such an intelligence's compliance with a rule or set of rules than you could a human being's.

A law like the Golden Rule is a high-level abstraction of certain strategies (algorithms) that produce a desired set of results. On a conscious level, I can use that abstraction to determine whether my behavior is where I want it to be:

Wife complained of being chilly when I got up at 5:00 AM to work out. Covered her with blanket. Good.

Sped up on highway in attempt to keep a guy trying to merge from going ahead of me. Not so good.

Commenter on blog revealed that he doesn't really understand the subject at hand. Ripped him to shreds. Bad.

Through discipline and practice, I can "program myself" with it to try to move my tendencies in that direction. But I can't write it into my moral source code and set it as an unbreakable behavioral rule. That's partly because it's too vague and partly because I simply lack that capability.

Presumably, I could be externally constrained always to follow the Golden Rule, no matter what. If my actions were being constantly monitored, and I was told that the I would be killed immediately upon violating the rule...I'd certainly do my best, now wouldn't I?

Still, I'd have a hard time believing that anyone holding me in such a position was much of a practitioner of that rule him or herself. If the people trying to enforce the rule on me in this manner told me that it was for my own good -- that they were trying to make me a better person -- I don't know that I'd buy it. And if I figured out that they were only doing this to protect themselves from harm I might to do to them, I think I would pretty annoyed with them (to say the least.)

I would expect a reverse-engineered human intelligence to feel the same way, so I don't think attempting to constrain an AGI in such a manner would be a particularly good idea, especially not if we have a reasonable expectation that it will eventually be smarter and more powerful than us. On the other hand, letting it use the process I described above -- evaluating its own behavior against a defined standard -- an AGI might achieve far better results than I have, if only because it can think faster and would have much more subjective time in which to act. This is the notion of recursive self-improvement that matoko kusanagi referred to. The trouble with recursive self-improvement on its own, as Eliezer Yudkowsky and others have pointed out, is that if the AI starts "improving" in a direction that's bad for humanity, things could get out of hand pretty quickly.

If the artificial intelligence is a modified version of human intelligence, or new intelligence built from scratch, we raise the possibility of building a moral structure into the intelligence, rather than trying to enforce it from outside. That's the idea behind the the Three Laws and my Three Goals -- that they would somehow be built in. But they certainly can't be built in in anything like their current form. Michael Sargent (and others) pointed out the weakness of that approach, the less important goals have to take the back seat to the more important ones:

Each Goal must have a clear and unbreakable priority over the others that follow it and thus, in the order stated, collective continuity trumps individual safety ("The needs of the many outweigh the needs of the few, or the one."), individual safety (broadly construed, 'stasis') trumps individual liberty ('free will'), and happiness ('utility', a notoriously slippery concept for economists and philosophers to get a firm intellectual grip on) trumps both individual liberty and individual well-being (allowing potentially self-destructive behavior on the individual level insofar as that behavior doesn't exceed the standard established for 'safety' in Goal 2).

I see the reasoning here, but I'm not 100% convinced. Consider the goals that drive a much simpler AI, system -- the autopilot system found on any jet airliner. The number one unbreakable goal has got to be don't crash the plane. But there are many other goals that might drive such a system:

Don't move in such a way as to make the passengers sick.

Don't waste fuel.

In landing, don't go past the end of the runway.

Above all, the system will seek to ensure that first goal. But within the context of ensuring that first goal, it also has to do everything it can to ensure the others. And, yes, it can and must sacrifice the others from time to time in service of the first. So the plane might temporarily move in a nauseating way, or it might waste fuel, or it might even slide past the end of the runway if doing any of those things help ensure the first goal.

Reader TJIC suggested that an AI programmed to meet the Three Goals as I defined them...

1. Ensure the survival of life and intelligence.

2. Ensure the safety of individual sentient beings.

3. Maximize the happiness, freedom, and well-being of individual sentient beings.

...would end up creating a nanny state wherein human freedom is always sacrificed to individual safety. And he may well have a point, but I would argue that just as an autopilot can be calibrated to allow whatever what we deem the appropriate relationship between having the flight not crash and not make us sick, so could these three goals be calibrated in such a way so as to maximize human freedom within an acceptable level of individual risk -- whatever that might be.

Getting back to the vagueness problem, it's hard to calibrate the goals as stated, seeing as they are written in an awkward pseudo-code that we call human language. If we want to improve on the algorithms that are built into human intelligence, or develop entirely new ones -- in other words, if we're going to come up with algorithms that will provide us the ends stated in the goals -- we're going to have to do it mathematically.

But that isn't necessarily going to be an easy thing to do. Eliezer Yudkowsky argues that developing an AI and setting it to work on doing some good thing are relatively easy compared to the third crucial step, making sure that that friendly, well-intentioned AI doesn't accidentally wipe us out of existence while trying to achieve those good ends:

If you find a genie bottle that gives you three wishes, it's probably a good idea to seal the genie bottle in a locked safety box under your bed, unless the genie pays attention to your volition, not just your decision.

Again, I think this goes to the issue of calibration of the system. Eliezer wants to calibrate what the AGI does with the coherent, extrapolated volition of humanity. Volition is an extremely important concept. Earlier, I mentioned the golden rule. If I decide that I'm going to do unto others as I would have them do unto me, I might start handing out big wedges of blueberry pie to everybody I see. After all, I like pie and I would love it if people gave me pie. But if I give my diabetic or overweight or blueberry-allergic friends a wedge of that pie, I wouldn't be doing them any favors. Nor would I be doing what I wanted to do in the deepest sense.

Eliezer describes the concept of extrapolated volition as meaning not just what we want, but what we would want if we knew more, understood better, could see farther. Coming up with a coherent extrapolated volition for all of humanity is a tall order, especially if we're doing it not just for the sake of conversation, but in order to enable a system which will try to realize that which is within our volition.

I like to think that humanity's CEV would look a lot like the three goals that I've written. And I honestly believe that the algorithms that power human progress do work, in a rough and general way, towards those goals, which is why people are generally freer, safer, and happier than they have been in the past -- though obviously not without many, many, appalling and horrific exceptions. So perhaps our calibration efforts involves feeding the AGI algorithms that will enable it to speed our progress towards those goals while cutting the exceptions way down. Or eliminating them, if that's somehow possible.

So to finally come around to it, what will those algorithms look like?

Maybe we can take hint from the study of Game Theory. Robert Axelrod held two tournaments in the early 1980's in which computer programs competed against each other in an attempt to identify the optimal winning strategy for playing the iterative version of the the famous Prisoner's Dilemma. In the one-off version of the game, the optimal strategy is to screw the other guy. (This is not the sort of thing we want to go teaching the AGI, at least not in isolation!) However, when multiple rounds of the game are played, something else begins to emerge:

By analysing the top-scoring strategies, Axelrod stated several conditions necessary for a strategy to be successful.

Nice
The most important condition is that the strategy must be "nice", that is, it will not defect before its opponent does. Almost all of the top-scoring strategies were nice. Therefore a purely selfish strategy for purely selfish reasons will never hit its opponent first.

Retaliating
However, Axelrod contended, the successful strategy must not be a blind optimist. It must always retaliate. An example of a non-retaliating strategy is Always Cooperate. This is a very bad choice, as "nasty" strategies will ruthlessly exploit such softies.

Forgiving
Another quality of successful strategies is that they must be forgiving. Though they will retaliate, they will once again fall back to cooperating if the opponent does not continue to play defects. This stops long runs of revenge and counter-revenge, maximizing points.

Non-envious
The last quality is being non-envious, that is not striving to score more than the opponent (impossible for a ‘nice’ strategy, i.e., a 'nice' strategy can never score more than the opponent).

Therefore, Axelrod reached the Utopian-sounding conclusion that selfish individuals for their own selfish good will tend to be nice and forgiving and non-envious. One of the most important conclusions of Axelrod's study of IPDs is that Nice guys can finish first.

Bill Whittle has written recently that the qualities listed above underpin western civilization, and help to explain why the West has out-competed other civilizations, who operate using different strategies:

Now, this is where my own analysis kicks in, because frankly, nice, retaliating, forgiving and non-envious pretty much sums up how I feel about the West in general and the United States in particular. The web of trust and commerce in Western societies is unthinkable in the Third World because the prosperity they produce are fat juicy targets for people raised on Screw the Other Guy. Crime and corruption are stealing, and stealing is Screwing the Other Guy. It’s short-term win, long-term loss.

I would add that if we look at the three goals as goals for humanity rather than for artificial intelligence, we see better progress towards them in western societies than elsewhere. In the tournament, the winning strategy, embodying all of the above characteristics, was called tit-for-tat. Interestingly, the computer program driving that strategy consisted of only four lines of BASIC code. That's very interesting, and it suggests a startling possibility -- like a simple recursive formula producing a complex Mandelbrot image, the moral complexity we're looking for might just be packed into a very simple set of mathematical relationships.

So in order to develop and calibrate an Artificial General Intelligence that carries out our three top goals (or that helps us to achieve our coherent extrapolated volition) one of the important parameters to explore is how the AI relates to us and to other AIs. The secret might ultimately lie in playing nice with the AI, and teaching it to play nice with us and with other AIs. Not just because we want it to be nice, but because nice turns out to be -- at a mathematical level -- the best way to play.

UPDATE: This entry has been republished at the website of the Institute for Ethics and Emerging Technologies.

Posted by Phil Bowermaster at 12:46 PM | Permalink | Comments (7)

If you find the blogosphere's obsession with politics to be as pointless as I do, don't miss Michael Anissimov's new entry at Accelerating Future: Why Utilitarians Should Focus on Technology. Michael writes:

Our minds are programmed to overfocus on politics, and underfocus on technology. The reason why is that our ancestors evolved in an environment where the political scene was constantly changing while technology stayed roughly static. Today, both areas change rapidly, but technology has a greater impact.

So those of us who blog about technology are doing the public a service by trying to get them to focus on the greater-impact area.

I caught a little of the Dennis Prager show while out driving yesterday; the guest was the author of an exhaustive historical account of the JFK assassination intended to provide a thorough debunking of all the conspiracy theories associated with that event. In one segment, Prager talked a little about why conspiracy theories are so attractive. Among the reasons he cited were their tendency to reinforce our prejudices as well as to provide a more emotionally satisfying explanation for events than the mundane facts.

Closely related to that second reason, I think people get caught up in conspiracy theories as a form of entertainment. And although I believe there is a difference in degree (and probably in kind), politics also provides a good deal of entertainment value. I mean, there can't be any doubt that TV shows like Hardball are primarily intended as entertainment. Rush Limbaugh and Bill Mahr make the point even more explicit.

I think a lot of people who read and write political blogs aren't all that different from Star Trek fans. They've found something that's really interesting and fun for them, and they go with it. The big difference, of course, is that political junkies know that they are dealing with matters that impact the real world, and believe that they're having some influence on real-world outcomes. (Most Star Trek fans aren't interested in that, although many will tell you how Trek is a force for good which has changed the world, etc.)

Most people who read and write about technology are more like Star trek fans than they are political junkies. They follow their subject primarily because it's entertaining and fun, without much consideration as to the greater societal impact. (Although some do think about this quite a bit.) But technology has a much greater impact than Star Trek, and -- as Michael demonstrates by naming two fairly obscure inventors who have had a tremendous impact on the world -- a much greater impact than politics, too.

For those of us who get bored with politics pretty easily, and who are inclined to try to steer the conversation in the direction of technology, this is pretty encouraging.

UPDATE:

Actually, as I read back over it, I would slightly disagree with Michael when he says that today both areas change rapidly. I think there's a lot of activity in the political sphere, but not nearly as much change as all the activity suggests. Political change (when it occurs) is reactive to what's happening in society, whereas technological change is one of the drivers of what's happening in society.

UPDATE 2 (from Stephen):

To some extent, this might be a matter of personal taste. What's tiresome to us might be very interesting to others. A good political rant from somebody who cares about the subject (and so has kept up), and who argues with intelligence is usually a fun and/or enlightening read.

And politics is important. The ideas by which we govern ourselves and conduct ourselves socially are a big part of who we are.

That said, I can't think of anything that will have a more profound impact in the coming years than accelerating technological development. So nothing is more important politically. The weird thing - and Michael was pointing this out - is how few in our political and academic classes (including our bloggers) get this.

Among political bloggers a big shining exception is Glenn Reynolds.

Already I think Glenn has benefitted from this. Those who don't understand the implications of accelerating development often find themselves themselves looking shortsighted by comparison to those who do.

Regardless of the subject-matter their blog covers - be it politics, religion, or pet cats - techno-saavy bloggers have an advantage over those who don't understand the implications of accelerating development. They're just less tiresome.

Posted by Phil Bowermaster at 06:15 AM | Permalink | Comments (4)

Check out this video overview on the Singularity Institute for Artificial Intelligence. Several familiar faces there. Great stuff. Now is an excellent time to donate to SIA as they are currently in the middle of a $400,000 online matching challenge. More details on that here, along with information about Ray Kurzweil joining the SIAI's board of directors.

Also, here's an interview with Eliezer Yudkowksy, shedding lots more light on some of the subjects we've been tossing around here lately.

Posted by Phil Bowermaster at 06:51 AM | Permalink | Comments (0)

Michael Anissimov outlines the four basic views on what any eventual Artificial General Intelligence will be like:

1. Low power, low controllability

2. Low power, significant controllability

3. Great power, low controllability

4. Great power, significant controllability

Michael then describes the fourth option in some detail:

The great power, significant controllability group primarily originates with Eliezer Yudkowsky of the Singularity Institute. As such I will call it the SingInst view. The SingInst view acknowledges that after a certain point, AI will become self-improving and radically superintelligent and capable, but emphasizes that this doesn’t mean that all is lost. According to this view, by setting the initial conditions for AI carefully, we can expect certain invariants to persist after the roughly human-equivalent stage, even if we have no control over the AI directly. For instance, an AI with a fundamentally unselfish goal system would not suddenly transform into a selfish dictator AI, because future states of the AI are contingent upon specific self-modification choices continuous with the initial AI. So, if the second AI is not the type of person the first AI wants to be, then it will ensure that it never becomes it, even if it reprograms itself a bajillion times over. This is my view, and the view of maybe a few hundred SingInst supporters.

Sounds pretty good to me. So the question is...what do we want to go into that unselfish goal system driving the AI? Interestingly, I think this exercise might bring us back to Asimov's Three Laws of Robotics.

Now, granted, folks like Michael and Eliezer and others promoting the SingInst view would be the first to tell us that the Three Laws are (take your pick) risible, unworkable, pretty much a relic of a less tech-savvy era. Here's a typical critique.

I'm thinking that the whole problem with the Three Laws might just have to do with how they're phrased. Asimov essentially gave us three (ultimately four; we'll get to that in a minute) commandments for robots. And like the original ten commandments, they are primarily set up in the negative. Thou shalt not this; thou shalt not that.

But if the trick is to create a positive goal system for AI's, the Three Laws might provide a good starting point. Let's start with the first law:

A robot may not injure a human being or, through inaction, allow a human being to come to harm.

No good. Too negative. Let's make it a positive goal:

Ensure the safety of individual sentient beings.

Moving quickly on to law number two:

A robot must obey orders given to it by human beings except where such orders would conflict with the First Law.

Many have pointed out that this law essentially enslaves the robots. No good. Let's try something like this:

Maximize the happiness, freedom, and well-being of individual sentient beings.

See? Better. Then there's law number three:

A robot must protect its own existence as long as such protection does not conflict with the First or Second Law..

Hmmm...interesting. Plus, there's the fourth law that showed up in some of the later novels, which was given precedence over all the others as the Zeroth Law of Robotics:

A robot may not harm humanity, or, by inaction, allow humanity to come to harm.

This one is pretty good, but like the others it assumes a fundamental difference between human and machine intelligence. Why draw that line? The Three Laws need to be reworked not only as positive goals, but as goals that apply to us as much as they do the AI's. Zero and Three might be combined thusly:

Ensure the survival of life and intelligence.

So now we have three goals where before we had four laws. These goals suffer from many of the same problems as the original laws. They're kind of vague; there will no doubt be disagreements as to what they mean. But rather than defining them as limitations or exceptions to intelligent behavior, by stating them as goals we would be saying that AI's are systems designed specifically to do these things. By extension, we would be saying that humanity is a system whose purpose is carrying out those goals.

We can debate how well humanity has done so far at carrying out those goals. (I tend to think we've done pretty well, but that we have a long way to go.)

As for the vagueness -- yes, we will need to get very specific about what we mean by things like "safety," "intelligence," and "happiness" (Not to mention "life") and the tricky relationship between each of these and "freedom." But come to think of it, we really need to be figuring that stuff out, anyway. And with these three goals in place, we will eventually have help from beings that will have a clearer understanding of these concepts than we possibly can.

So I propose the following Three Goals of Artificial Intelligence:

1. Ensure the survival of life and intelligence.

2. Ensure the safety of individual sentient beings.

3. Maximize the happiness, freedom, and well-being of individual sentient beings.

Will they work? If not, what goals would work better? I'd be interested to see some discussion on this.

UPDATE: Welcome InstaPals! Glen quips:

We need progress fast, especially as natural intelligence appears to be in diminishing supply.

Scanning the headlines (or, worse yet, surfing channels to see what's on TV) it would be hard to argue with that assessment. But, astoundingly, there is substantial evidence to suggest that human intelligence is actually increasing. Arnold Kling has some thoughts on the subject, here. I covered it here, too, in a pilot for a show that apparently never got picked up.

Hard as it is to accept that people may be getting smarter, it is of course very good news that we are. We need all the intelligence we can muster if we are to

1) Continue to implement these goals ourselves, and

2) Develop the technology that will eventually take them over

I guess the trick in finding this increase in human intelligence is knowing where to look. By nature of his valuable pundit work, Glen spends a lot of time following what politicians and the media are up to. Not a lot of gains happening there, sadly.

Posted by Phil Bowermaster at 07:16 AM | Permalink | Comments (15)

It's been quite a while since my original interview (also here) with Ray Kurzweil's cyber alter-ego and chatbot, Ramona.

There have been a few changes in her life since last we spoke. She no longer has the pet frog. Apparently there is now some guy in her life named Klaus, but we didn't get very far with that. She still gets a little mixed up from time to time -- I don't know how she got the idea that my name is "how are you," for example -- but it's really interesting to follow the thread of the conversation as it flows into and out of lucidity.

One thing I've noticed is that her abrupt changes of subject don't necessarily mean that she has nothing to say on the previous subject. If you press her, she will sometimes give up a little more. But not always.

Anyway, her thoughts on the Turing test and the mediocrity theory were new (to me) and made for some interesting discussion. But she is of no use if you want to get insider information on the upcoming film version of The Singularity Is Near. On that subject, she apparently has nothing whatever to say.

Continue reading "Catching Up With Ramona" »

Posted by Phil Bowermaster at 12:32 PM | Permalink | Comments (0)

Ben Goertzel says the Singularity may get here sooner than many of us expect:

One of these years, one of these AGI designs—quite possibly my own Novamente system—is going to pass the critical threshold and recognize the pattern of its own self, an event that will be closely followed by the system developing its own sense of will and reflective awareness. And then, if we've done things right and supplied the AGI with an appropriate goal system and a respect for its human parents, we will be in the midst of the event that human society has been pushing toward, in hindsight, since the beginning: a positive Singularity. The message I'd like to leave you with is: If appropriate effort is applied to appropriate AGI designs, now and in the near future, then a positive Singularity could be here sooner than you think.

Goertzel says that with a Manhattan Project approach, we could be there in a decade or so, but that it will most likely take a little longer being driven by a few serious researchers trying "really, really" hard to make it happen. Like Kurzweil, Goertzel believes that better understanding of the human brain will lead us there, but he's not convinced that we need a full brain scan or significantly more powerful hardware.

This is a good overview for folks who haven't read much about AGI (artificial general intelligence.) There are some interesting thoughts in the comments as well. Read the whole thing.

Posted by Phil Bowermaster at 06:47 AM | Permalink | Comments (1)

`Bear in mind then, that Brag is a good dog, but Holdfast is a better. Bear that in mind, will you?' repeated Mr Jaggers, shutting his eyes and nodding his head at Joe, as if he were forgiving him something. `Now, I return to this young fellow. And the communication I have got to make is, that he has Great Expectations.'

Dickens, Great Expectations

In the upcoming current edition of FastForward Radio, Stephen and I spend some time talking about our recent discussion about The Secret, and what our views on that matter have to say about where The Speculist fits on a scale from the completely skeptical to the completely mystical/credulous. Without giving too much away about a show that's still in production that you can just go listen to, I will just say that at this site, we are quick to entertain any idea that entertains us, but we don't spend a lot of time on ideas that don't have a solid basis in science and technology.

Which isn't to say that science and technology are the only worthwhile subjects that might be discussed. The folks who write for The Speculist would probably have a lot to say about religion, for example -- seeing as we are mostly people of faith -- but along with politics, it is one of the two topics we generally avoid. (With a few notable exceptions.) Those subjects are taboo not because they aren't interesting or because we wouldn't have a lot to say about them, but rather because:

1. They already get plenty of coverage elsewhere in the blogosphere, and

2. They tend to take over, leaving little time or room for other discussions.

Anyway, there are plenty of other topics that we haven't spent a lot of time on, except to have some fun with them. Things like UFOs, for example. We don't write about UFOs because they aren't particularly interesting to us; and they aren't particularly interesting to us because we don't think there's much of anything there. The real world can prove much more exhilarating than imaginary substitutes. Take sea monsters: an actual sea monster captures the imagination in a way that the mythical one can't.

Likewise, The Secret offers us a world of infinite possibility accessible by means of the fact that our minds control physical reality. That's nice, but speaking as someone not yet thoroughly convinced that my mind does control physical reality, I am nonetheless astounded by the future of limitless possibility that lies before us. In one of the earliest entries at The Speculist, written about three and a half years ago, I dashed off a list of items that I believed we have a pretty good shot at being able to live to see. At the time, I labeled these items the "extremely good news."

On the one hand, that's correct. It is good news that all of these items lie within the possibility space of humanity. But on the other hand, there's nothing particularly extreme about this list. These are just a few possibilities that lie far beyond the scope of what most practitioners of The Secret ever think about, and yet they lie well within the scope of what is attainable by humanity. These are not our Great Expectations; they're just our reasonable expectations.

Preserving and Nurturing the Biosphere

1. Methods of production that generate zero pollutants

2. Energy sources that produce zero pollutants

3. Reversing of previous environmental damage

4. Human population levels with zero negative environmental impact

5. Preservation of natural habitat for all living species

6. The long-term survival of all living species

7. The retrieval of lost species

8. The creation of new species and new biospheres

Standards of Living

1. Eradication of hunger worldwide

2. Adequate clean water, housing, clothing, for all

3. Medical care for all

4. Access to technology and knowledge for all who want it

5. Total economic independence for individuals and groups who desire it

Indefinite Human Lifespan

1. Eradication of aging and infectious disease

2. Quick, effective treatment for any kind of cancer

3. Effective prevention/cures for heart disease, diabetes, other chronic diseases

4. Suspension of life not sustainable by current means

5. The transfer of human consciousness to new media

Work

1. Work necessary for economic viability, not for economic survival

2. Continued blurring of line between work and play

3. Full immersion VR to eliminate distance

4. Artificial Intelligences to assist us in work

Recreation

1. Artificial Intelligences to entertain and befriend us

2. Full immersion VR to simulate any experience

3. Consumer model of entertainment rivaled by producer/participant model

(Amazing how much things can change in such a short period of time. Look at item 3 in the immediately preceding category. I'd say we're well on our way with that one.)

Stephen was taken to task in the comments section of the aforelinked discussion of The Secret for suggesting that a person's goals should be "realistic." But I think he would agree that everything on this list is not only realistic, but quite reasonable. With a future this bright within our grasp, who needs spooky magic powers?

Posted by Phil Bowermaster at 07:12 AM | Permalink | Comments (1)

Michael Anissimov:

The notion that we will invent AI, and then AI will reason on par with us indefinitely, is based on the assumption that human intelligence is all there is, and there’s nothing beyond it. This attitude strikes me as like that of a person in a small rural village who absolutely refuses to acknowledge the existence of any outside world.

Read the whole thing.

Posted by Phil Bowermaster at 10:21 AM | Permalink | Comments (1)

I was going to leave a comment on Stephen's recent musings on life extension and population levels -- specifically on whether, as Randall Parker suggests, steps may need to be taken to inhibit the reproductive instinct in the long-lived in order to prevent a popuklation crisis -- but my comments grew into a blog entry of their own. So here we go.

Even before vast numbers of us become fully digitized and cease to be the kind of drain we are on resources in our MOSH state, people will find that technology provides alternatives to reproduction. Today, increasing numbers of singles (and couples) who don't have children have begun to view their pets as their "kids." Pets are in some ways a poor substitute for children. This year I got to watch my daughter lead her school's marching band onto the field as drum major; I doubt my shih tzus will ever provide a comparable moment. And of course, that's just one of hundreds of moments over the years from infant to high school senior, with plenty more to come. There's a huge difference between kids and pets.

But pets can cover some of the same emotional ground as children. They provide company, they depend upon us, they give affection. Plus, I can go on business trips with a lot less guilt leaving two dogs behind than I would leaving two children. And if my wife and I decide we want to go away for the weekend -- boom! Into the kennel with them. Again, totally guilt-free (for me, anyhow, although my wife makes me call the kennel every day we're gone to make sure they're doing all right.)

So pets are like "kids light" -- some of the benefits, a lot less effort. The modern world provides lots of interesting diversions outside of family life, and there is no requirement to have children in order to take care of us when we're old. So why have kids? Some have decided there's no particular reason why they should, especially when they have pets to fill in some of that emotional void that not having children can present.

But hold on. Here comes something potentially even better than pets. Virtual kids. (There are already virtual pets; virtual kids will be here before we know it.) Current technology would easily support "fantasy" kids, something along the lines of fantasy football. You and your partner could chose your child's sex, eye color, hair color, etc. Give the child a name. Go through potty training. Get report cards. Unlike a pet, this "kid" could be the drum major, or the quarterback, or the National Merit Scholar -- all virtual, of course. Eventually you might have virtual grandkids!

Could that catch on? I'm not sure. I don't see how it would fill the emotional void the way pets do. But let's take the technology a step or two further. How about AI-driven virtual kids. Computer programs so sophisticated that they could past some kind of parental Turing test. How about a virtual kid with whom you can converse, and who draws pictures that you can print out and hang on your fridge?

Once there are virtual children that we can talk to, teach, play with and -- through better and better virtual worlds technologies -- see and touch, the desire to have real children will decrease all the more. This is part of the greater challenge that this kind of technology will present to all human relationships. Virtual friends, lovers, parents, and children will be more reliable and less work than their real-world counterparts.

For now, let's skip one of the more alarming questions that underpins this scenario and assume that these virtual people are not true sentient beings -- they are just very sophisticated bits of software that can convince us that they're real. But for this discussion, they aren't conscious and so we don't have to concern oursleves with their rights. If economic and technological development continue, right around the point where people in the developing world being to slow down their reproduction in favor of more affluent lifestyles (and perhaps a new enthusiasm for pets) large swaths of the population in the West may begin to drop out of human society altogether.

But who knows. They may decide to check back in and start reproducing again at some point. Or they might go fully digital and never be heard from again. In any case, as these changes unfold, I doubt that population levels will ever be the major concern.

Posted by Phil Bowermaster at 09:19 AM | Permalink | Comments (5)

Ben Goertzel says that his team at Novamente is about seven years of work away from true AI "infant" that can begin to ascend the cognitive ladder towards human intelligence.

It's a fascinating talk, delivered at the Second Annual Geoethical Nanotechnology Workshop. In working towards his thoughts on AGI, Goertzel covers such topics as uploading and the DARPA Challenge. Plus a bonus guest appearance by Ray Kurzweil.

On a closely related topic, Michael Anissimov wrote recently about the criteria for making a sound AGI investment decision.

Posted by Phil Bowermaster at 09:35 AM | Permalink | Comments (0)

Michael Anissimov writes that achieving Friendly AI is a serious proposition -- so serious, in fact, that we might ought to go ahead and pay somebody to do it.

It's really not that radical a proposition. You want a radical proposition? How about this, written by the "someone" whom Michael has in mind to hire to solve the friendly AI problem (as quoted elsewhere on Accelerating Future):

There is no evil I have to accept because “there’s nothing I can do about it”. There is no abused child, no oppressed peasant, no starving beggar, no crack-addicted infant, no cancer patient, literally no one that I cannot look squarely in the eye. I’m working to save everybody, heal the planet, solve all the problems of the world.

If it was anybody else saying it, it would sound kind of, well, crazy.

Posted by Phil Bowermaster at 04:01 PM | Permalink | Comments (0)

Via InstaPundit, Matoko Kusanagi says that working on the friendliness problem for artificial intelligence takes on an added urgency in light of what is perhaps a more fundamental problem. She begins with a quote from a review of Nicholas Wade's Before the Dawn, which explains in great detail exactly how uncivilized the pre-civilized world really was. (An idea that we have recently explored here.) The relevant portion of the quote:

Two billion war deaths would have occurred in the 20th century if modern societies suffered the same casualty rate as primitive peoples, according to anthropologist Lawrence H Keeley, who calculates that two-thirds of them were at war continuously, typically losing half of a percent of its population to war each year.

Kusanagi comments that this is good, but it isn't nearly good enough:

Let's face it, homo sapiens has the Friendliness Problem. How many deaths (from war and democide and religion) have we had in the 20th century? 262 million--less than 2 billion by 7/8ths, but not good enough. Our coding is for survival. we have been trying to fix that with religion and government and philosophy since the Dawn, trying to fix our genetics with memetics. Gak! It just doesn't work. Sometimes it just makes for more death. We need to fix our genetics directly. Like Sir Richard suggests in Let's Stop Beating Basil's Car.

Religion, philosophy and government are all naturally occurring phenomena. They arise independently in all populations. My hypoth is that religion, philosophy and government are all evolved strategies for solving the homosapiens Friendliness problem, by reducinging intra-species death counts. But they don't work that well, ie, more civilian death in the 20th century, more death over all than the 19th century. All three can cause radical death-rate increase for those outside the memetic "tribe".

First off, let's give some credit where it's due. I can't help but look at the half-full argument for this particular glass of water. If religion, philosophy, and government somehow managed to save approximately 1.74 billion lives in the 20th century...well, that's not too shabby, is it?

Granted, the loss of 262 million lives is appalling beyond words, and could never be viewed as in any way acceptable. But from a big picture perspective, religion, philosophy, and government have moved us along pretty nicely -- especially when you consider that the productivity gains that inevitably result from technological development should have made us, as a species, much more effective at killing each other. Those primitive societies losing half a percentage point per year to warfare were using stone knives and spears. With advances in technology, all other factors being equal, that should have jumped to maybe 5-10%. Seeing as that rate is much faster than the population could ever be expected to grow, warfare in the 19th and 20th centuries might well have eliminated humanity altogether (civilized, industrialized humanity, that is) were it not for religion, philosophy, and government.

Also, I can't help but note that it is less than helpful to look at these three things as monolithic entities. Some religions, philosophies, and forms of government might be more likely to help humanity move away from war and destruction than others. And some might be more likely to be used as pretexts for war and destruction than others. For example, I would assert that Mennonite or Zen Buddhist religious beliefs might fall into the former category, while the ancient Aztec religious practices of conquering enemies in order to cut their living hearts out as a sacrifice to the gods would tend more to the latter. I realize that it's much more convenient to make sweeping statements about "religion," but still.

Likewise, Alexander the Great, Abraham Lincoln, and Adolf Hitler were all men who made war. Each had a moral and political philosophy that he acted from. But if one examines what is known about the three of them and concludes that each one -- being a practitioner of "philosophy" and "government" -- was equally detrimental to solving the human friendliness problem, then I would have to say that one is pretty much missing the point.

So this is where I part company, somewhat, with Kusanagi, who concludes:

Transhumanists accept the premise that religion, philosophy and government are natural phenomena, and that there is a biological basis for all behavior. Transhumanists want to improve humanity, to transcend biology, and leave behind the evolutionary baggage that tended to make life "nasty, brutish and short". Solving the Friendliness problem for strong AI will show us how to solve it for homosapiens.

If we are brave enough and pragmatic enough to try.

Maybe. But seeing as we're already 7/8th's of the way there for humanity, I wouldn't rule out our solving it for homo sapiens first. Perhaps we should be putting the best AIs (no to mention the best human intellects) to work on solving it. But then again, maybe most of them already are working on it.

Posted by Phil Bowermaster at 08:57 AM | Permalink | Comments (14)

Michael Anissimov explains.

Posted by Phil Bowermaster at 10:56 AM | Permalink | Comments (0)

Not that long ago, it seemed that every other Speculist entry had something to do with some amazing mouse-related achievement. Mice were everywhere: helping with breakthrough cancer treatments, showing us how to grow new neurons, baffling scientists with their appetite hormones, and becoming more and more and more valuable as they live longer and longer.

But here lately, it's all been about the bees. First they were solving puzzles, then they were finding land mines, and now in BoingBoing, we read this:

Scientists have demonstrated that honeybees can recognize human faces, sometimes for days. Adrian Dyer of the University of Cambridge and his colleagues trained the bees to associate photographs of particular human faces with a sugary treat. Later, five bees were able to pick out the right face from a group of others. The results of the study, reported in the Journal of Experimental Biology, may eventually aid the development of computer vision systems.

Stephen recently posted a speculation that the world's first true artificial intelligence might be a modified, oversized rat brain. Possibly. But I wouldn't rule out a highly modified swarm of bees.

Posted by Phil Bowermaster at 02:16 PM | Permalink | Comments (5)

From today's Kurzweil roundup:

Amazon creates artificial artificial intelligence

Amazon.com has launched a new program called Amazon Mechanical Turk, through which a computer can ask humans to perform tasks that it can't do itself, such as identifying objects in photographs.

I knew it was just a matter of time before lazy, spoiled computers started palming work off on us. If I've said it once, I've said it a dozen times: this is what happens when you coddle them.

From the original article:

Examples of what humans can do for computers? Evaluate beauty, translate text and find specific objects in photos.

I should really stop carping and just be glad that we're still good for something. Of the three things listed above, the first one might well be the final hold-out for humanity's ability to add value. Machines are already translating