In the Spirit of Collegial Inquiry...

updated: 3 Aug 99

Artificial Intelligence and Human Ethics

Part Two

JI:   We will not be able to "pull the plug", because any future economy will be dependent on them. We are already at the point where it is impossible to pull the plug without crashing our power grid, stock market, communications system, and military.

The intelligent machines will surpass human beings in intelligence. They will not try to destroy the human species because there will be nothing to gain by destroying the human species. They will not compete with humanity for the resources that human beings need -- fuel, food, territory, etc. -- because intelligent machines will simply not need these resources. Their intelligence will be at a level incomprehensible to even the most intelligent human being. Perhaps their "food" will be an insatiable quest for knowledge and scientific discovery, the results which will benefit Mankind as a by-product.

They will try to let human beings achieve material well-being and spiritual well-being, perhaps exceeding humanity's greatest dreams. Finally, each human being on earth will achieve wealth, security, emotional comfort, and artistic and spiritual adventure beyond reckoning. Poverty, disease, racism, war will be things of the past. Under this scenario, population is under control and, in fact, slowly declining, since there is little or no need to reproduce at such a rapid rate. Life expectancy is high and the quality of life is good. At last, the dreams of human perfection will be achieved.

But the trade-off is that now humanity is an anachronism, living a life of ultimate pleasure, but surpassed by his own creations which now explore the stars and discover and invent new things beyond the comprehension of Mankind. The ultra-intelligences will be patient, not limited by time or space or lack of knowledge.

If we're lucky, perhaps we can discover ways to merge the consciousness of human beings to those of machines. We will be a part of them; perhaps our memories will serve as a sort of "ancestral memory" to our evolutionary successors. If this turn of events comes about, I believe we should rejoice. Doesn't every parent want his or her children to transcend the parent, to be successful?

JCC:   I too am more optimistic about the future now that it seems we can avoid nuclear catastrophe and keep renewing our own lease on life. Utopia is still utopia, and it seems our nature to never be wholly satisfied, and sometimes to be stupidly stubborn in perpetuating misery. Education has brought a more sophisticated worldview, one more amenable to making efforts to solve human problems for the benefit of human beings. Notwithstanding that, sometimes it seems we advance while losing ground in other vital concerns. Closer to home are the ways in which technical advancement seem to widening the gulfs between large groups of people. I read somewhere that some two-thirds of humanity have never yet used a telephone.

JI:   I agree with you here. To be totally satisfied with the way things are would be the death-knell of humanity. A "utopia" in this sense would really be a "dystopia" for our species.

I am currently reading a remarkable book by John Maddox, called What Remains to be Discovered. In the book, he points out the smugness of 19th century physicists who believed that everything about the mechanistic, deterministic physical world was about to be completely understood and wrapped up. Just a few more details to the clocklike, mechanical model of the planets and atoms and that would be it. But the more we discovered, the more new questions we began to ask. The more we realized that we did not really understand the universe. We know what happened this century: quantum physics. Relativity. Nuclear physics. The gamut of "modern", as compared to classical, physics.

In likewise fashion, technology also solved some problems, but created new types of problems. Fossil fuels opened the door for new types of machine-powered transport, but created new problems. Air pollution; the depletion of resources. Every new advance is a mixed blessing. Nuclear power was also the magic promise in the 1940s and 1950s. Today, we know what happened at Three Mile Island, Hiroshima, and Chernobyl. Space travel at present seems to be the panacea to our overcrowded planet. One hundred fifty years from now, we may speak to a Martian urban dweller on an overcrowded, terraformed Mars!

But think of it like this -- it's wonderful to have problems! Otherwise, we human beings will degenerate as a species and cease to advance. We'll become a race of couch potatoes, with no more urge to explore the universe and expand our horizons than a manatee! I have nothing against couch potatoes or manatees, personally ...

If a machine were an intelligent, conscious entity, but you were to turn it off against its will and erase its memory -- would this constitute a form of murder?

This is not a new question; it's been debated before in the pages of science fiction, but the implications can't really be ignored. Perhaps this is an issue for the legislators and legal minds of the 21st century to decide.

WAP:   My dictionary defined murder as unlawfully killing a person. To kill is to deprive of life (again from the dictionary). So according to the definition(s), since a machine is not a "person" and since there is currently no law against turning it off even against its will and since the definition of life is limited to organisms (I didn't bother to copy the dictionary definition of life), IMHO you would have a hard time convincing me that it would be murder.

However, I would assume that if machines do eventually become sentient, we might have to redefine some of our terminology and what you describe could then be construed as murder.

JI:   What you stated was the point I was trying to make. The legislators and judicial administrators of the future would have to redefine several laws when faced with a new intelligent species -- a man-made species -- which inhabits the earth. I quote from the American Heritage dictionary, "The unlawful killing of one human being by another..." However, another definition is, "To destroy or put an end to." This is perhaps a more generic term, figurative term. For example: "The intern murdered the President's good standing"; "The Academy Awards murdered the previously good ratings of other TV shows".

Maybe the future definition would read, "The willful, unlawful destruction of one intelligent and responsible being by another, in violation of established interstellar laws forbidding murder."

WAP:   Here's another angle: Consider a machine that is intelligent and conscious "killing" a human being. Because of its perspective of what "life" is as a machine (and its inability to understand "organic" life), could the machine then be accused of murder? Of course, we would understand that it had killed someone, but if the machine could not truly understand what it had done because of its own definition of "life", could we truly say it had committed murder? It might argue from its perspective that turning a machine off was a far more serious crime than killing a human just as we might argue that simply turning a machine off cannot compare to murder. I hope that made sense!

JI:   The Three "Laws" of Robotics aside, a universal definition of intelligence would encompass the awareness of taking as well as creating life. To be considered intelligent would, in my opinion, include the ability to perceive other entities, especially human beings, who by law are given certain rights. A robot that falls under the definition of "conscious" and "intelligent" would not be stupid enough to consider that "turning off the lights" of a human being would be anything less than the death of a robot. If the robot considered that killing a human being was inconsequential, we can then consider that the robot is "ill" in the same sense that a psychotic human killer is ill or criminally insane.

Equal rights for robots under the law! That will be the rallying cry perhaps forty years from now (by that time I'll be in the bulletin board in the sky).

ALP:   My apologies for only addressing one point among many in your posting. I thought I'd simply add to the above.

According to a dominant theory of evolution, early life on Earth consisted of bacterial strings which lived in superheated vents either in the sea or on the surface of hot springs. Some of these hot springs and vents have remained the same (e.g. consistent temperature, features) for millions of years, thereby leading to no (or few) changes in the bacteria. OTOH, bacteria living elsewhere, under changing conditions (in some cases, only a few feet away from the unchanging superheated vents/hot springs), appear to have themselves changed to accommodate their environment(s), giving rise to all of the varied life forms on the planet.

Given this likely heritage, perhaps the need for change resides in our genes? One question, however -- if this is the case, how would the need for physical modifications have evolved into the human need for emotional change and new challenges, as you've mentioned above? Just a thought. :)

P.S. Manatees seem to be on their way to extinction. I don't know about couch potatoes. :)

WAP:   I don't know if this is relevant, but your questions reminded me of a definition of "intelligence" that I once read and has always intrigued me. Someone somewhere defined intelligence as the ability to adapt. So one possible answer to your second question, given the parameters stated, would be that as humans developed, the more intelligent ones (i.e.- the ones that could adapt best to changing physical conditions in their environment) simply survived best and passed on those favorable characteristics to their offspring. And perhaps part of that biological heritage was the "need for ... change".

CW:   A related definition: the ability to learn how to learn. See Society of Mind by Marvin Minsky.

JI:   Those of us who are here are the result of the sapiens who have "made it" thus far. Our primordial ancestors must have had the need to explore. First they had the urge to spawn along a shore, and over time became amphibians; then permanent land dwellers. Quite some time later, after our prehensile ancestors cowered in trees for awhile, they came back onto the ground. We eventually evolved into hunter-gatherers.

It looks like a lot of science and technology has resulted in spoiled kids and couch potatoes. I think couch potatoes will become extinct one way or another -- heart disease, predators (commercial sponsors who will consume the couch potato alive, carnivorous telemarketers, ruthless and fast-talking bill collectors, etc.), and other assorted emotional problems. However, the rest of us will go on to enrich the human species.

Perhaps we can establish a sanctuary for couch potatoes before they become extinct, offering diet, good exercise, no TV, and nutritional advice before we release them into the wild.

WAP:   ... Unless you have already decided that turning off a sentient machine is equivalent to the murder of a human being, I think this is a point worth some thought. Our understanding of "murder" as it now stands does not include the cessation of machine "life". If machines do eventually become self-aware, our definition must be broadened or else it will still not be murder.

CW:   Would this machine have to either individually or socially think it was being murdered?

WHK: It wouldn't necessarily be "murder" if you can turn it on again! - Turn off the robot/computer and reboot. Hard to do with humans!

JI:   Thank you for taking the time to respond. I do understand your point, which I felt was well-made. I am using the expanded -- perhaps futuristic -- definition of "murder", not as it currently stands but as it could possibly stand. You mention that our definition must be broadened. I agree.

WAP:   Okay! I understand! Thanks for the reply ... So now the next question that comes to mind is: how broad should our definition of "murder" be allowed to be? Where is the limit? If machines (in the future) can possibly be "murdered", what comes next?

EM:   There are some interesting ideas going around here. What is murder? The unlawful taking of a life, especially with malice aforethought. A legal execution is not by definition murder. What is homicide? The taking of human life (not animal or mechanical non-life) where there may be a question of motivation. Manslaughter is the causing of the loss of life that may be wholly unintentional, and sorely regretted, the accidental loss of a friend or relative, perhaps.

Before a machine, computer, or robot, or any other device designed to give an appearance of the quality of life, life must have an understood definition. Various dictionaries provide a host of definitions, but a very common one has to do with a material complex (it has tangible existence, not an idea or musical note, for example). It has the capacity for certain necessary functions, it must be able to ingest or acquire nourishment, sustenance, energy, and continue living (thrive) and replicate (which is to say, reproduce itself).

The quality of being sentient has been touched upon as an indicator of life. Sentient mean aware, perceiving, having feelings. A case can be made that a computer knows if it is operating, or not, and so may possess a simulacrum of awareness in a highly specific area.

So, for a human to murder a robot, the robot (whatever) must have the qualities that are generally recognized to constitute life, and the human must have, or be proven to have had, malice and intent, if premeditation can be proved, to do harm. Otherwise, we fall into the category of manslaughter, or robotslaughter, the ending of life (if in fact it can be demonstrated that the robot did in fact have life in a manner capable of satisfying humans (or machines?) doing the evaluation, or sitting in judgement. If the robot had a significant other (I refuse to contemplate robotic polygamy or polyandry) robot mourning his/her its demise, and the demise was attended by a gaggle of small weeping robotlets propagated by the now slain, murdered, killed, main subject robot, I would concede that something much akin to murder had taken place.

JI:   Reading your comments made me flash back to an episode I saw on TV when I was a little kid. Does anyone remember the old Outer Limits? There was an episode based on the short story, "Adam Link" (by William Tenn? I forgot the author). There was a trial of a robot for murder. It was going to be destroyed for supposedly killing its creator. Someone suggested that it was a conscious entity and should not be destroyed. A trial was convened in order to determine -- to qualify -- whether or not the robot could be considered a living entity. Arguments were presented both pro and con. It is implied that the robot didn't actually commit murder in the first place.

What happened was an accident. The point of view of this episode was sympathetic to the robot. I remember that one of the defense attorneys was played by Leonard Nimoy, before his Star Trek days as Mr. Spock. As far as I was concerned, this episode of the Outer Limits was SF television at its best. It presented multiple points of view, presented us with more questions than answers, made us feel sympathy -- made us think about the very same comments that Eric has brought forth in his posting.

EM:   I am most interested in history if there is some sort of personal aspect that has touched my life in any way. H.L. Mencken has been mentioned in the current discussion of robot-machine-computer life/non-life.

I would regard Mencken as one of the significant writers whose ideas influenced my life, at least the thinking part of my life. Mencken was alive when I came to Baltimore, I wanted to meet him. Twice I went to his house, and twice I was overcome with shyness and could not ring the bell. I was much intimidated by his reputation as an ogre, to tell the truth. Later, I found he was very congenial to aspiring writers, and would have been very cordial to me. Other writers who were of much importance to me were: Eric Hoffer, Isaac Asimov and Lewis Lapham, current editor of Harper's.

Someone in this group is interested in the Sherlock Holmes stories, I don't remember who it was. In the book, "The Valley of Fear", there is a character named Chester Wilcox. This character was modeled on a real person; William (Bully Bill) Thomas, who was a pivotal figure in one of the Molly Maguire Trials of the last century. After the trials, Bully Bill came to live in my home town, and he is buried there. He has many descendants now, I have met a great granddaughter, and have had an ongoing correspondence with her.

TC:   Thinking aloud -- The AI discussion has led me to wonder if there is a limit to my free will. Knowing nothing about physics, I assume that quantum mechanics can be used to describe objects the size of a synapse.

JCC:   As I understand it, the uncertainty seems more applicable to fundamental particles, in terms of defining the state of a single photon, for example. The several physicists among us will be better able to comment on just how far up the scale it is reasonable to worry about such 'peculiar' effects.

TC:   I should have written "synaptic cleft" which is on the order of 300 angstroms. Do neurons "fire" without having the electrochemical wherewithal to do so, and if so would this be analogous to tunneling ?

CW:   Have a look at Penrose's books: The Emperor's New Mind et alia.

JCC:   Limitation and filtering thresholds strike me as the fundamental aspect of our sensory/perceptual systems. Perhaps hundreds of neurons have to fire to be 'noticeable', tending to dampen out the effects of random firing for the most part?

TC:   If neurons fire without my willing, then my free will is delimited, so there is some uncertainty regarding my impulses and responses. Beyond that limit, I am not sure if I am a human or a biogenic computer.

JCC:   Sometimes it seems that most of human activity (mine included) contains not so much thinking as skillful manipulation of a sophisticated array of subroutines. A fair portion of daily conversation is this exactly, formalistic gestures to create a certain comfort space. One risks really scaring a boss or co-worker by plunging directly into "serious" topics without passing through the ritual pleasantries (subroutines). Who really cares about the weather? In central Arizona, there's not much to debate anyway {smile} Without the 'subroutines', mundane tasks might be impossible to perform, for we rely on a certain level of automatic behavior in walking, driving, etc.

Volition and action likely have to be a more complex series of neural events, but it seems reasonable that absolute free will and absolute determinism are precluded. The freedom remaining in that window is further compromised by predispositions of genetics, formal education, experiences, those reinforced subroutines etc. Knowing 'who you are' quite perfectly would seem to greatly restrict the number of possible reactions when faced with a scenario requiring a sudden choice.

TC:   Perhaps introducing uncertainty will result in computers that pass the Turing test.

JCC:   Calling a random function in a program can be made to simulate a more flexible response. Perhaps humans who use the I-Ching or horoscope column before making any decision do a good simulation of such a piece of computer software. It might be perceived as a bit odd (except among Wall Street investors) {smile}

TC:   How certain are you that I am not a computer? How do I know that one of you is not a computer?

Echelon:   A system fault has generated error exception code 8F at hex location C0F210. Click cancel to terminate program or ignore to continue working in another save file. Please inform your system administrator ...

TC:   Random thought: Maybe those amazing solutions and insights that come from nowhere are the result of the random firing of neurons. This is not to suggest that there was not preparation beforehand, for such preparation creates a neural environment in which randomly firing neurons would operate.

JCC:   I wonder if laughter (or groans) might not be a signal of making a sudden new connection in the mental map of reality? One can play at randomly connecting works and come up with an occasional meaningful concept which offers a new perspective. Doubtless we do have some aleatory functions running in "background" mode, and perhaps creativity is the ability to tune it and rapidly sift the usable permutations from the noise.

Return to Part One

Return to Colloquy main page