I. Existing technologies in artificial intelligence Artificial Intelligence (AI) is an experimental science. Its task is the re-establishment of reasonable discourses and actions by means of artificial devices. Applied aspect of artificial intelligence involves the tasks solved by computer without explicit algorithmic solutions and sometimes with unclear objectives. Moreover, it is often used “human” methods regarding how to deal with such problems. The scientific aspect of artificial intelligence concerns the explanation of the natural intelligence and simulation of human tasks. The field of artificial intelligence has begun to develop after the appearance of computers. Today, the elements of artificial intelligence are used in many spheres from household appliances to the control of nuclear power plants. The development of artificial intelligence is parallel to the acceleration of computers and advances in cognitive science. It is expected that after several decades artificial intelligence will reach human level and then exceed it.
A. speaking computers: A.L.I.C.E
The first robot companion was created already in 1966. Joseph Weizenbaum program Eliza imitated dialogue with a psychotherapist. In the conversation Eliza played a role of active listener though she showed off her cleverness. The basic principle of the Eliza was the paraphrase of interlocutor’s replicas. According to Joseph Weizenbaum “ELIZA performs best when its human correspondent is initially instructed to “talk” to it, via the typewriter of course, just as one would to a psychiatrist. This mode of conversation was chosen because the psychiatric interview is one of the few examples of categorized dyadic natural language communication in which one of the participating pair is free to assume the pose of knowing almost nothing of the real world” (Eliza (elizabot.js)). To make sure that this is true I visited the following website http://www.masswerk.at/elizabot/ and had the following conversation with Eliza: ELIZA: “Is something troubling you?” ME: “Yes, I have a headache.” ELIZA: “I understand.” Me: “Can you help me?” ELIZA: “Do you wish to believe I help you?” Me: “Yes.” ELIZA: “You are sure.” And so on.
In 1995, scientists started to develop other virtual interlocutor such as A.L.I.C.E. (The Artificial Linguistic Internet Computer Entity) program. This virtual chatty girl compares phrases, said by person, with the replicas in its knowledge base. Today, A.L.I.C.E. is considered to be one of the best of its kind. However, neither A.L.I.C.E. nor Eliza was able to pass Turing test in which the interrogator has to identify who is his or her interlocutor: a person or a computer. However, A.L.I.C.E. became the winner of the Loebner Prize three times in 2000, 2001, and 2004 (“AI people behind and bidding for the Loebner Prize”). Since 1990, Loebner Prize is annually awarded the winner of the content, in which computer programs compete in Turing test (“What is the Loebner Prize?”).
B. Robots that can carry out tasks like ADAM.
For the first time in the history, the scientific discovery was made by a laboratory robot. It is a robot Adam, created by scientists from Aberystwyth and Cambridge Universities. He put forward a hypothesis about the genes and enzymes of yeast and conducted experiments for verification. As a result, a number of simple but useful discoveries were made, which then were confirmed by researchers. The team leader of Adam’s designers, Professor Ross King, even expressed the hope that one day robots and people will work in laboratories side by side, “Ultimately we hope to have teams of human and robot scientists working together in laboratories. Because biological organisms are so complex it is important that the details of biological experiments are recorded in great detail. This is difficult and irksome for human scientists, but easy for Robot Scientists” (“Robot scientist becomes first machine to discover new knowledge”).
Adam was aimed with a colossal database on the biology of yeast, automatic devices for experiments, stocks of cell culture and chemicals as well as with a powerful artificial intelligence (Bland). According to Professor Ross King, Adam was the first model of its kind, and now scientists have created a more sophisticated robot named Eve, who, along with Adam will search for new drugs against malaria and other tropical diseases. Professor Ross King also said that “If science was more efficient it would be better placed to help solve society’s problems. One way to make science more efficient is through automation. Automation was the driving force behind much of the 19th and 20th century progress, and this is likely to continue” (“Robot scientist becomes first machine to discover new scientific knowledge”).
C. Other artificially smart systems
It should be said that scientists have long tried to create robots that could replace people in certain professions. Robots are increasingly used in medicine. In some areas they are able to work more efficiently, with greater accuracy and less likelihood of error than the doctor. Soon, it will be possible to combine the robo-surgery with diagnosis technologies (expert systems have long been used for diagnosis, X-rays analysis, etc.). Surgical system, such as da Vinci has already been put into action. According to Dr. William E. Kelley Jr, the operating surgeon, “Complex operations that are now done through large incisions, and some operations that we can’t even envision today, will become routine through tiny incisions with this system in the years ahead” (“FDA OKs Robo-Surgery”).
Considerable part of robots used now has the embryo of artificial intelligence. They can navigate a bit in the environment as well they can recognize objects. In 2005, five of twenty three automated test vehicles successfully completed the course of over 132-mile Mojave Desert (“Innovations that reflect our customers’ mobility aspirations”, 5). “Eighteen failed to complete the course because of mechanical failures or sensor problems. The robotic vehicles had to navigate a course designed to mimic driving conditions in Iraq and Afghanistan” (Chang).
Planes can already fulfill the entire flight from takeoff till landing flying by the autopilot. One of the British company is working on artificial intelligence system in order to extend the flight range, “The on-board system being developed by Roke Manor Research uses video feeds to collect information about clouds, ground surface conditions and other elements, then crunches the data to develop flight paths that steer gliders toward rising air”(Demerjian). In Japan, France, and other countries, there are automatic trains that use artificial intelligence to make the trip more comfortable for passengers (Bertram 1).
II. Limitations of existing artificial intelligence machines
A. Limited activities.
There are many problems related to the term “Artificial Intelligence”. The scope of problems is sufficiently broad and rather uncertain. In the most general sense, artificial intelligence is a combination of methods and tools aimed to process information (knowledge) in accordance with the experiences gained in the period of learning and adaptation in dealing with a variety of intellectual tasks.
Works regarding the creation of artificial intelligence systems are now in the initial stage. Basically, such systems exist in the form of specialized computer programs that can solve in isolation intellectual tasks such as playing chess or other intellectual games, composing music, proving of mathematical theorems, the dialogue with a person in a language close to the ordinary English, Japan and others, medical diagnostics, and etc. In recent times, it have appeared artificial intelligence systems designed in the form of specialized equipment such as image recognition systems, speech recognition systems, radar signals recognition system, and etc. (Tugui).
B. Maximum Program
The main difficulty in designing such systems lies not in finding the elements for their creation but rather in finding the logic for the joint work of large number of such elements. Features of artificial intelligence system are determined rather by the algorithms and programs installed in it than by its technical realization. It has to be emphasized that the capacity of artificial intelligence (as well as ability of the human intelligence) fundamentally is limited because the amount of memory, speed of reading and storing information in artificial intelligence systems are practically always limited. It is evident especially in solving intellectual tasks of high-dimension (Cognitive Science).
The objective of research (the maximum program) is the creation of artificial intelligent systems that can perform no worse (and possibly better) than human the work that are traditionally considered to be the sphere of intellectual labor. However, it is necessary, above all, to separate many intellectual tasks into two classes: the creative tasks (when there are no method of solving them at the time of their appearance), and routine tasks (the solution of such problems is known, you only need to re-find a solution).
C. More Questions that Answers
However, the modern computer is doomed to act only by means of commands written in the form of the programs. For this reason, it copes well with the routine intellectual process, often even better than person, but it is not able to find a solution to the problem that not previously described in its program. Therefore, any task with which the artificial intelligence can not cope at the moment, appears as, in essence, indeed intellectual, creative task. However, as soon as a way of problem solution is formalized into program, it goes into the category of routine tasks.
Thus, the artificial intelligence is artificial but not intelligence. This self-negation is partially removed by the redefining of artificial intelligence (the minimum program) as a field of research aimed to ensure that computers can perform the functions that are currently difficult for them, and especially those that only people can perform.
The purpose of the industrial robots is to create machines that are not worse by flexibility than people. Methods of scene analysis and recognition of images that is another line of artificial intelligence, allow to construct devices that are capable to collect details of the mechanisms that lie somewhere as a heap in the store, just as the person used to do it (Minsky).
Nevertheless, even here, there are more questions than answers in spite of the well-known achievements. It is very difficult for robots to do things that people easily perform. Thus, there is no trouble for the people to cross the street or to take a fragile object in the hands such as an egg, not breaking it. Such things are a mere trifle for a human, but the computational component of the robot is not yet in a position to do this “trifle” without outside help.
III. What could be
A. Automatic weapons
Automatic weapons have appeared during the mechanization of mass production in the XIX century. Military experiments with vehicles moving without a driver (including even radio controlled airplane) began in the First World War.
Two events related to the same problem are always a reason for more careful consideration. There was an international conference “Future Force 2020: Are We Prepared?” organized by the Netherlands Defence Academy and the TNO (The Netherlands Organization of Applied Scientific Research) in The Hague in March (“Robots wars in 2020”). At the same time, in the United States it had just issued the book “Wired for war: the robotics revolution and conflict in the twenty-first century”, the author of which is Singer P. W., security analyst at the Brookings Institution and director of the 21st Century Defense Initiative (Singer, “Author Biography”).
In order the acute actuality of the topic of military robots became quite obvious, it is enough to draw attention to the fact that to-date more than forty states are building robots for military use.
Unmanned aircraft, or simply Raven, is the most popular robot aircraft in the U.S. Army, has a length just thirty-eight inches long and weighs four pounds, so it is launched from the hand. It can be thrown up and forward like a spear. In the air, Raven can be up to half an hour, flying at low altitudes (four hundred feet) and carrying on board three cameras, one of which is infrared (“RQ-11B Raven Small Unmanned Aircraft System”). Soldiers like this model because it allows seeing what is happening over the next hills or assessing the situation in nearby city blocks. The most importantly is that soldiers can launch and land spy as well as program the route, i.e. it is not necessary to ask the superiors to send aircraft for help. Almost anyone can learn how to manage such Raven, so that for the first two years of the Iraq war, the number of “Raven” in the troops increased from twenty five up to eight hundred (Singer, “Military Robots and The Laws of War”).
B. Pseudo Humans
Most people think that with the creation of an electronic brain will start a paradise for humanity. Robots managed by electronic brain will work without rest, creating an abundance of wealth. Humanity will spend time on fun, entertainment, art, commanding the robots, or, at the best, will be engaged in creative work. It is a regular and deep confusion. There was no any case in history and will never be a situation where the highest level of intelligence was a servant at a lower level. The whole world of microorganisms, plants, and animals are our ancestors. However, whether we ever have been a servant at our closest congener as apes? No, we do not think about such situation at all. Man is ready to accept equality in rights the other person that is the essence of the same mental level, but he will never accept equality in rights any monkeys. Furthermore, we have no remorse in breeding useful for us animals, killing them for food when we want without a mercy. We destroy noxious plants and microorganisms. And the closest congener monkeys are used for medical experiments and test new drugs. This is despite the fact that we all belong to one biological formation.
So, how other “civilization” built on a completely different formation that is on far more advanced electronic principle will treat humanity? In the same manner because we belong to the lower levels: to use us when we are useful, and to destroy, when we disturb. With the creation of an electronic brain will begin the end of human civilization. Our exclusion from the reservation is likely to be gradual, but short. Although E-brains may do something at the beginning in order to calm the unsatisfied, attract the rulers and the inhabitants to their side (Bostrom “How Long Before Superintelligence?”).
C. Robots and Morality
In addition to the aforementioned, there is another view on what is happening. Behind the Earth’s civilization is a shameful history. How many times humanity thoughtlessly made genie out of the bottle whether bombs, poisonous gases, or nuclear weapons, even before there were carefully reviewed terrible consequences of the use of destructive technologies (Bostrom “Existential Risks: Analyzing Human Extinction Scenarios and Related Hazards”). Today, there is a very real chance to take account of past mistakes. So, talking about combat robots and morality, it would be useful to recall the famous Isaac Asimov’s Three Laws of Robotics:
- “A robot may not injure a human being or, through inaction, allow a human being to come to harm.
- A robot must obey orders given it by human beings except where such orders would conflict with the First Law.
- A robot must protect its own existence as long as such protection does not conflict with the First or Second Law” ( Asimov).
Today, specialists who create robots did not remember the “old-fashioned and naïve” ideas of Isaac Asimov. However, if they suddenly recall, they do it only to indicate that all of his stories from the series “I am a Robot”, in its essence, demonstrate the internal contradictions and limitations of the fundamental laws.
However, no one wants to see a very significant thing that for Isaac Asimov it was initially unacceptable the idea of autonomous armed robots, programmed for the intentional infliction of harm to humans, not to mention the murders. In Asimov’s future world this could not be in principle. Thus, if in our today’s world everything is differently, it means that there is a big difficulty with the understanding of the human life value, and hence with the foundations of morality in modern society. The society that is not considered murder as immoral is terrible ill. However, it is not clear how to cure it.
- “AI people behind and bidding for the Loebner Prize”. Compute Scotland, 2008. 7 May 2009 http://www.computescotland.com/1774.php
- Asimov, Isaac. “Three Laws of Robotics”. 7 May 2009 http://www.auburn.edu/~vestmon/robotics.html
- Bertram, Volker. “Cyber-Ships – Artificial Intelligence Technologies for Ships”. ENSIETA. 7 May 2009 http://www.skibstekniskselskab.dk/public/dokumenter/Skibsteknisk/Download%20materiale/2008/10%20marts%2008/cyberships_DTU.pdf
- Bland, Eric. “Self-directed robot scientist makes discovery”. MSNBC.com, 2009. 7 May 2009 http://www.msnbc.msn.com/id/30016175
- Bostrom, N. “Existential Risks: Analyzing Human Extinction Scenarios and Related Hazards”. Journal of Evolution and Technology, 2002. 7 May 2009 http://www.nickbostrom.com/existential/risks.html
- Bostrom, N. “How Long Before Superintelligence?” International Journal of Futures Studies, 1998. 7 May 2009 http://www.nickbostrom.com/superintelligence.html
- Chang, Alicia. “Stanford Wins Pentagon Robot Race”. Washingtonpost.com, 2005. 7 May 2009 http://www.washingtonpost.com/wp-dyn/content/article/2005/10/09/AR2005100901086.html
- Cognitive Science. AAAI.org, 2008. 7 May 2009 http://www.aaai.org/AITopics/pmwiki/pmwiki.php/AITopics/Co gnitiveScience Demerjian, Dave. “Artificial Intelligence Gives Gliders a Lift”. Wired.com, 2008. 7 May 2009 http://www.wired.com/autopia/2008/08/artificial-inte/
- Eliza (elizabot.js). Norbert Landsteiner, 2005. 7 May 2009 http://www.masswerk.at/elizabot/ –
- “FDA OKs Robo-Surgery”. Wired News Report, 2000. 7 May 2009 http://www.wired.com/science/discoveries/news/2000/07/37536 –
- “Innovations that reflect our customers’ mobility aspirations”. 7 May 2009 http://gb.volkswagen.com/2005/fileadmin/g_bericht/pdf/en/GroupTopics_en.pdf
- Minsky, Marvin. “The Age of Intelligent Machines: Thoughts About Artificial Intelligence”. KurzweilAI.net. 7 May 2009 http://www.kurzweilai.net/meme/frame.html?main=/articles/art0100.html
- “Robot scientist becomes first machine to discover new scientific knowledge”. University of Cambridge, 2009. 7 May 2009 http://www.admin.cam.ac.uk/news/dp/2009040201
- “Robots wars in 2020.” Media Center of the University of Sheffield, 2009. 7 May 2009 http://www.sheffield.ac.uk/mediacentre/2009/1188.html
- “RQ-11B Raven Small Unmanned Aircraft System”. 7 May 2009 http://www.af.mil/factsheets/factsheet.asp?id=10446 Singer, Peter Warren. “Author Biography”. Wired for War. 7 May 2009 http://wiredforwar.pwsinger.com/index.php?option=com_content&view=article&id=49&Itemid=56
- Singer, Peter Warren. “Military Robots and the Laws of War” The New Atlantis, Number 23, 2009, pp. 25-45. 7 May 2009 http://www.thenewatlantis.com/publications/military-robots-and-the-laws-of-war
- Tugui, Alexandru. “Reflections on the Limits of Artificial Intelligence”. Association for Computing Machinery. 7 May 2009 http://www.acm.org/ubiquity/views/v5i38_tugui.html?searchterm=artificial+intelligence
- “What is the Loebner Prize?” Home Page of The Loebner Prize in Artificial Intelligence, 2007. 7 May 2009 http://www.loebner.net/Prizef/loebner-prize.html