Of all the technological developments of the modern age, none is as complex as artificial intelligence. The idea that a non-human, manufactured entity could advance to a point of emulating human behavior is enough to make people shiver. The idea of something so advanced that we will not be able to tell the difference may be even worse. In 1961 Philip Dick wrote Do Androids Dream of Electric Sheep?. It portrayed a world with these exact characteristics. Artificial Intelligence had advanced to a point of near-perfect human replication. In creating this world, Dick was echoing the speculations and fears of a.i. that had build up by that time. It would then continue to influence those opinions in the years after.
It is necessary to look at the development of artificial intelligence in order to put this idea into context. The concept of intelligent and aware constructs began to emerge in the 1950s and 60s as several scientists in many fields came together to discuss the possibilities of advanced computer research. The first major step was a scientific conference at Dartmouth College in 1956. Here, the general concepts and possible paths of research for a.i. were fleshed out. As described in Artificial Intelligence: A Modern Approach, this conference was “the birth of artificial intelligence.” This was mostly a theoretical stage yet attending experts predicted that with a huge investment, working technology could be available in a generation (16). After being officially established, a.i. research and discovery exploded. Computer programs, a brand new idea, were already conquering algebra problems and speech recognition. Some could even reproduce English (18). It was clear that artificial intelligence research was going to be at the forefront of scientific efforts.
Two influential experiments defined the entire foundation of intelligence research. The first was the Turing test. Introduced in 1950, this test sought to test a machine's ability to demonstrate human intelligence (McCorduck 95). This test features three subjects: a human, a computer, and an interrogator. Each subject is place in separate rooms, but are able to communicate with each other freely. During the test, the interrogator asks a series of questions to determine which is the artificial subject. Because of the limited technology of the time, communication is limited to text transmission. The Turing test was far from perfect and two major concerns came to light during the process: human intelligence vs general intelligence and real intelligence vs simulated intelligence (Saygin &ump; Cicekli). The first weakness stems from the reality that human behavior can be unintelligent and intelligent behavior can be inhuman. The second weakness is a result of the likelihood that a machine can be taught to emulate intelligence rather than create it. This “synthetic intelligence” is not the same genuine human ability (227-258). The androids of Do Androids Dream of Electric Sheep? Seemed to have passed this synthetic phase. Their motivations are much more than programing to look human. This is clearly evident with their preparations against, and fears over the systematic killing of other androids on earth. The group being focused on in the book show clear self-preservation sentiments, an undoubtedly organic characteristic.
The very first truly artificial intelligence was the Logic Theorist in 1955-56 (167). Designed by Allen Newell, Herbert Simon, and JC Shaw with the objective of creating a program that could prove theorems using goal-oriented behavior and rule application. This was an attempt at having a computer that could mimic the problem solving skills of humans. In its first official test, the Logic Theorist was able to correctly prove thirty eight of the fifty two theorems presented (168). One stunning achievement of the presentation was when the intelligence provided a proof more elegant and effective than the one submitted by Newell, Simon, and Shaw.
With the creation of artificial intelligent programs comes the ever-imminent consequences. People aren't likely to accept a new generation of beings superior to themselves. As humanity 's advancement approaches total artificial sapience, we realize the possible implications. Do Androids Dream of Electric Sheep? was the first work to popularize artificial intelligence and start the general population thinking about it. The central idea on a.i. came also with a question of what it meant to be human. Could intelligence advance so far as to allow other beings to be “human?” Once manufactured constructs have human organs, human emotions, and human cognition, is there anything else? Is it possible to progress beyond that? With such advanced creations coming into existence there are going to be significant ramifications. Addressing these probable consequences may help us understand unforeseen problems that are going to happen.
Oxford University looked in depth at the most significant of these implications of advanced intelligence work. Lead expert Nick Bostrom rated the human-machine competition as one of the most direct consequences. One significant fear that develops out of this is the idea that human-level artificial intelligence leads quickly to greater-than-human-level artificial intelligence. The scientific community is quite unlikely to stop their work when they reach a point where human-quality machines are achieved. Such a discovery would be an astounding feat. With an achievement so admirable, it may seem like further advancement is simply impossible. The fallacy is that by stopping with human-level intelligence, the scientific community would be ignoring what makes artificial intelligence such a complex and notable field. One of the original concepts established by the Dartmouth Conference was that artificial intelligence may be able to surpass human capability and result in a construct of yet unparalleled potential. If it is possible to make computers as capable as humans, it should theoretically be possible to make them better. It can be predicted, with a high degree of likelihood, that the latter is gong to happen. Although the time-scale is obviously uncertain, progress may even develop faster that we think. Artificial Intelligence research is occurring in conjunction with the research into the structure of general technology. More advanced hardware technology directly increases the potential of constructs. Bostrom go so far as to theorize that humans and machines are not likely to be matched for long. “Shortly thereafter, humans will be unable to compete intellectually with artificial minds” (12).
Androids comments on this by way of the Rosen Association. It is in competition with other massive manufacturing firms like the Grozzi Corporation over the production of androids ad other artificial technology. There is an ever present demand for a better and more capable model of android. This demand is so high that without androids, the entire economy and society of the space colonization effort would collapse. As a result, each successive generation of artificial intelligence is more advanced than the previous. The NEXUS-6 is just the most recent in the years of this demand-fueled advancement. Although still incapable of empathy, a critical component of humanity, it resembled humans in every way down to blood, organs, and behavior. This new model was even capable of superhuman ability. Speed, strength, and intelligence easily outstripped the NEXUS's human competition. It can be inferred that the corporations would continue to build more advanced models to satisfy demands as time continued. Eldon Rosen summarized the very basic principle of android advancement:
“We produced what the colonists wanted. We honored the time-honored principle underlying every commercial venture. If our firm hadn't made these progressively more human types, other firms in the field would have. (54)”
The market is the true authority over what does or doesn't get done in any corporate environment. In the environment of Do Androids Dream of Electric Sheep?, the market says that androids will become better as long as the people demand it. With the total reliance on android technology in the colonies, this demand shows no signs of falling.
The other significant implication of artificial intelligence was described by Bostrom as: “Unlike other technologies, artificial intelligences are not merely tools. They are potentially independent agents (16).” He claims that it would be a mistake to see intelligent machines as tools or as some kind of disposable labor. While it may be possible for humans to create machines for specific uses or to only think along a narrow set of parameters, this is not the limit of what is likely to be available. Our constructs are likely to be able to become capable of “independent initiative” and of making their own plans. If this were to occur, it may be more appropriate to view these intellects as persons rather than machines. Expanding on this, Bostrom hypothesized that by trying to impose controls on the intellect's motivations, we would in effect create a class of highly effective “slaves (17).” Slavery, carrying strong social stigma, may result in an intense political and ethical battle. Is it right to force a being to do as commanded when that being is capable of its own independent and unique thought?
Dick saw this dilemma as well. The androids had been used as slave labor in the colonization of space. They were used to do the hazardous work humans were incapable of. In addition to harsh manual labor, the androids were forced into servant positions. As an incentive to leave earth for the colonies, each human is promised an android house butler. The ethical problems had either been resolved or never occurred in the first place. It is possible that because each android was successively more advanced, the progress towards true artificial intelligence was more subtle. Humans didn’t notice as its servants became more aware of their situation. It was this awareness that started the bounty hunter program. As androids begin to realize the extent of their exploitation, they flee the colonies. This flight inevitably takes the androids to Earth. The humans create a bounty hunting initiative to destroy any intellect legally living there.
Current progress on artificial intelligence is not yet near the “flesh and blood” robot phase. There are subtle advances towards that milestone though. Computers are already at a stage of understanding speech in nearly every spoken language. As work continues, we will be ale to achieve more and more natural-sounding voices and pitches. The most recent example of this is Apple's SIRI. This personal assistant responds to spoken questions with an appropriate answer. Anything from directions to the mall to the meaning of life is fair game. Even the popular search engines are a form of artificial intelligence. Computer systems are easily able to read through web pages, extract the relevant information, and create an accurate and concise report of what the user likely wants. Computer intelligences are always advancing. Each stage of development increases our reliance and this in turn promotes advancement.
Do Androids Dream of Electric Sheep described a dark world where artificial intelligences and humanity came into a form of conflict. Its predictions and ramifications are not easily dismissed and hold true even today. We are already seeing a.i. surpass humans in some fields. IBM's WATSON has passed humanity in terms of intellect. It will be some time before it is self-aware of that fact, but it is a milestone nonetheless. We also have artificial “slaves” in some small sense of the word. Machines to much of the heavy labor previously done by man. Massive assembly plants now only employ a few technicians to watch overt the computers. While a far cry away from slavery, the process has begun. We are addicted too we usefulness of machines and the addiction is impossible to cure.