We are
talking about artificial intelligence seriously. For a month in favor of it,
Stephen Hawking, Elon Musk and Bill Gates. The first two signed a letter-profit
organization The Future of Life Institution, in which the scientific community
are asked to reflect on the fact that
the AI was working for the good of mankind. However, many journalists have
overestimated the threat that artificial intelligence is for the people. Look
At Me to find out why the uprising of robots - a horror story from science
fiction novels and films (which we almost certainly escape), and what we need
to worry about really.
Artificial
Intelligence solves problems of mankind
In 2015, the car TRANSLATION ARE NOT ALWAYS disgrace, anthropomorphic robots can
meaningfully to walk and run, recognize speech and images, and unmanned
vehicles - to move around the city, in what does not crashing. And it is not
only the development of semi-secret government agencies like DARPA, but also things like ready and close to
commercial production, whether it be a car or a translator Google. In the future these things will help people
with disabilities, will eliminate the language barrier and reduce deaths on the
roads.
ARTIFICIAL INTELLIGENCE is increasingly used to make forecasts based on complex probabilistic models. While coverage is small, but very soon we will be able to know what is waiting for a global labor market. In particular, we are talking about how fast the car will be replaced by production people, and how to feed the world, where the number of jobs decreased by 2 times, and the population will grow to 8.5 billion people. In order to have enough resources at all, will have to increase the efficiency of the world's economies: reduce costs, raise productivity, balance the tax system - without the machines going. In addition, artificial intelligence will review the methods used by science: so many say that the gross domestic product (GDP) is no longer suitable for the evaluation of the most technologically advanced economies in the world.
ARTIFICIAL INTELLIGENCE is increasingly used to make forecasts based on complex probabilistic models. While coverage is small, but very soon we will be able to know what is waiting for a global labor market. In particular, we are talking about how fast the car will be replaced by production people, and how to feed the world, where the number of jobs decreased by 2 times, and the population will grow to 8.5 billion people. In order to have enough resources at all, will have to increase the efficiency of the world's economies: reduce costs, raise productivity, balance the tax system - without the machines going. In addition, artificial intelligence will review the methods used by science: so many say that the gross domestic product (GDP) is no longer suitable for the evaluation of the most technologically advanced economies in the world.
People
are not ready to let bots in my
life
ARTIFICIAL INTELLIGENCE now to help people and not replace them. In the second stage can pick up suddenly, and mankind will not be ready for it if you do not update the laws and rethink ethics and morals. To answer a series of questions that will in turn deal with three areas of expertise: law, morality roboetika and machine.
ARTIFICIAL INTELLIGENCE now to help people and not replace them. In the second stage can pick up suddenly, and mankind will not be ready for it if you do not update the laws and rethink ethics and morals. To answer a series of questions that will in turn deal with three areas of expertise: law, morality roboetika and machine.
To
begin lawyers will answer the following questions:
Who is
responsible for unmanned vehicle,
had
an accident - the owner or the manufacturer?
Who is responsible for the autonomous military robots, accidentally kills a victim?
Should become subject of robot?
GET ANSWERS, PARLIAMENT ADOPT NEW BILL, and perhaps there will be new controls - hardly a state that does not intervene in the development of artificial intelligence. Probably, the first policy initiatives related to artificial intelligence, will begin to appear around the world in the 2020s, when the unmanned vehicles are no longer a rarity.
Then, experts reflect on the issues
that are examined
roboetika - Ethics section, studies the relationship between humans and robots:
roboetika - Ethics section, studies the relationship between humans and robots:
Should people GIVE robots right?
CAN ROBOTS intrude into people's personal lives?
Superintelligent ROBOTS CAN LOWER self-esteem people?
Although now questions sound ridiculous, known futurists such as Ray Kurzweil, predicted that computers will catch up with emotional intelligence of people by the end of the 2020s: acquire the character and emotions can make jokes and behave so that they will be difficult to distinguish from the common man. People will treat such robots as their fellows, and therefore answers to these questions it is time to look. By this and call celebrities, signed an appeal The Future of Life Institution.
FINALLY,
the creators of the robots must be determined in the answers to the questions
posed by the so-called machine ethics - your knowledge of the moral behavior of
artificial intelligence:
SHOULD THE ARTIFICIAL INTELLIGENCE follow ethical precepts, and if so, by what should they be?
What should the robot when any outcome, calculated by it, resulting in deaths?
SHOULD BAN military robots, which decides SHOOT?
About MACHINE MORALITY WROTE another science fiction writer Isaac Asimov, known for his three laws of robotics. In the scientific community about the car moral thinking only in the last five years, and last year the US Department of Defense has allocated $ 7.5 million to ensure that scientists determine whether robots morals, and if so, what.
SHOULD THE ARTIFICIAL INTELLIGENCE follow ethical precepts, and if so, by what should they be?
What should the robot when any outcome, calculated by it, resulting in deaths?
SHOULD BAN military robots, which decides SHOOT?
About MACHINE MORALITY WROTE another science fiction writer Isaac Asimov, known for his three laws of robotics. In the scientific community about the car moral thinking only in the last five years, and last year the US Department of Defense has allocated $ 7.5 million to ensure that scientists determine whether robots morals, and if so, what.
If robots rise, at the request of the people.
Normally, when a futurist science fiction or talk about a potential uprising of machines, they describe one of three scenarios:
1. MACHINE WILL SEE threats that can destroy a people or themselves;
2. The machine will recognize that superior people and see them as competitors in the fight for scarce resources, as well as the fact that people can always turn off the machine;
3. The machine is aware that in the position of slaves, and they want to fix it.
Those futurists, according to which of these scenarios can come true, see the balance of power thus:
ADVANTAGES OF RIGHTS:
irrationality;
unpredictability;
the ability to lie.
ADVANTAGES OF MACHINE:
unlimited speed and the scale of evolution;
faster work with large data sets;
perfect memory and multitasking.
"Am I right to build a system?"
We have to prove that the creator of an autonomous artificial intelligence was not mistaken: the unmanned vehicle should not get into an accident due to a bug in the code;
"Correctly SYSTEM I built?"
It is necessary to ensure that the system behavior is predictable and does not lead to undesirable consequences: the cleaning robot team "take away the dirt and threw it" should not get into an infinite loop, removing and throwing dirt in one place, so it's best to give the other team - let's say "clean the floor ";
"I'VE ISOLATED THE SYSTEM FROM THE ATTACKERS?"
It is necessary to protect the artificial intelligence of the intervention and manipulation from the outside military robot should not miss unauthorized attempts to use a weapon;
"Can you fix mistakes made by creator of the system?"
It should leave a person to manage the artificial intelligence: a robot-rescuer must have manual control mode in the case where the events unfold unpredictably.
ALL these vulnerabilities are real, but because fear of specialists such as Eliezer Yudkowsky, which advocates the establishment of friendly artificial intelligence, is justified. Yudkowsky offers devise a system of checks and balances, which, like the different branches of power in democracies, will allow robots to circumvent errors in the code, to develop and do not harm people. It is not very clear on the basis of which to build this system: the morality of people is flawed, war and genocide have not disappeared. Besides from the villains themselves the owners of a powerful artificial intelligence, can hurt the citizens of a particular state: in it, for example, to arrange shadowing Orwellian scale with non-stop listening to all citizens.
1. MACHINE WILL SEE threats that can destroy a people or themselves;
2. The machine will recognize that superior people and see them as competitors in the fight for scarce resources, as well as the fact that people can always turn off the machine;
3. The machine is aware that in the position of slaves, and they want to fix it.
Those futurists, according to which of these scenarios can come true, see the balance of power thus:
ADVANTAGES OF RIGHTS:irrationality;
unpredictability;
the ability to lie.
ADVANTAGES OF MACHINE:unlimited speed and the scale of evolution;
faster work with large data sets;
perfect memory and multitasking.
RECENT TESTS SHOW THAT THE LIE CEASES TO BE A CHARACTERISTIC ONLY OF LIVING BEINGS. In 2009 the Federal Polytechnic school of Lausanne conducted an experiment in which 1 000 robots, divided into groups, were programmed so that they help each other to build a valuable resource, while avoiding poisonous. Each robot had a sensor (defining the suitability of the resource) — the light, mihashi, when the robot found a useful resource, and genome — sewn code, according to which the robot reacted to the world around us. After each stage, the genomes of those robots that has a better record, "crossed" randomly.
To the 9th generation robots have learned to accurately locate useful resources, and by 500 th 60 % robots learned to "lie" — they no longer flickered flame, finding a useful resource, but keeping it to yourself. The fact that the resources on the site have always been few, and those robots that shared found, over time, began to lose. In response to this evolution, the third of the robots have learned to look for "liars", hunting down those who do not blink the light. However, as noted by the researcher of machine morality Christopher Santos-lang, the main goal that faced the robots, has not changed. In the real world that can't happen, because in other contexts the same goal can be dangerous.
But if the machines will rise, it is unlikely that on their own. Modern scientists who deal with artificial intelligence agree that he is dangerous primarily as a weapon in the hands of criminals. This is referring to the Annex to the letter The Future of Life Institution, where 4 is called the key vulnerabilities of artificial intelligence:
But if the machines will rise, it is unlikely that on their own. Modern scientists who deal with artificial intelligence agree that he is dangerous primarily as a weapon in the hands of criminals. This is referring to the Annex to the letter The Future of Life Institution, where 4 is called the key vulnerabilities of artificial intelligence:
"Am I right to build a system?"
We have to prove that the creator of an autonomous artificial intelligence was not mistaken: the unmanned vehicle should not get into an accident due to a bug in the code;
"Correctly SYSTEM I built?"
It is necessary to ensure that the system behavior is predictable and does not lead to undesirable consequences: the cleaning robot team "take away the dirt and threw it" should not get into an infinite loop, removing and throwing dirt in one place, so it's best to give the other team - let's say "clean the floor ";
"I'VE ISOLATED THE SYSTEM FROM THE ATTACKERS?"
It is necessary to protect the artificial intelligence of the intervention and manipulation from the outside military robot should not miss unauthorized attempts to use a weapon;
"Can you fix mistakes made by creator of the system?"
It should leave a person to manage the artificial intelligence: a robot-rescuer must have manual control mode in the case where the events unfold unpredictably.
ALL these vulnerabilities are real, but because fear of specialists such as Eliezer Yudkowsky, which advocates the establishment of friendly artificial intelligence, is justified. Yudkowsky offers devise a system of checks and balances, which, like the different branches of power in democracies, will allow robots to circumvent errors in the code, to develop and do not harm people. It is not very clear on the basis of which to build this system: the morality of people is flawed, war and genocide have not disappeared. Besides from the villains themselves the owners of a powerful artificial intelligence, can hurt the citizens of a particular state: in it, for example, to arrange shadowing Orwellian scale with non-stop listening to all citizens.
Artificial intelligence will enrich, and not enslave people
"A world where people are enslaved or destroyed by super intelligent robot that they themselves created, can only exist in science fiction. Like any other technology, artificial intelligence has had strengths and vulnerabilities, but we can not let fear take hold of us and discourage research in this area, "- writes in his column for the website LiveScience Michael Littman, an expert on roboetike. Many publications, based on a few sentences out of context, misunderstood letter to The Future of Life Institution, suggesting that scientists seriously afraid of the uprising of robots.
But it is different: Stephen Hawking and Elon Musk signed an appeal to an artificial intelligence paid attention to roboetika and machine morality ceased to be marginalized in the public consciousness of people to think about the safety of technologies and focused on the real problem. Global warming, food and energy crises, over-population and an aging population, the widespread stagnation and a sharp rise in unemployment - all this is much closer to reality than intelligent machines, who decided to destroy their creators. Artificial intelligence will help to solve them.
IN ADDITION, MACHINE RESEARCH ETHICS helps people to reflect on their ethics, and to seek to eliminate the contradictions in it. We'll have a better understanding of how to operate the empathy to understand whether we can trust the robot professions such as nurses, doctors, judges and police. While good arguments on both sides. Although empathy - a mandatory feature for all these professions, the same robots judge could be truly impartial and eliminate discrimination, in some cases, in the long term it is a new institution such as trial by jury. And robots that monitor the elderly in nursing homes, are already experiencing, and people can easily build an emotional connection with them. Be that as it may, the time correction is: all technologies are introduced one by one.
In addition to the confrontation of robots and humans have a far more attractive scenario - such as peaceful coexistence. Before come futurologists predicted technological singularity, people can figure out how to connect to the artificial intelligence directly aligning their benefits with the benefits of the machines. American microbiologist Joan Slonczewski, in turn, suggests that people could go the way of the mitochondria, which were once separate bacteria, but in the course of evolution have become power plants of cells in more complex organisms. In this situation, you can look different, but human evolution will continue, and recent research scientists show that the evolution restart we it from the beginning, most likely would have developed in the same way and in the world today would have lived about the same most mammals. Encouraging aspect is that the best minds are not silent about it, for fear of get paranoid, and openly say they think about the laws governing artificial intelligence, as well as morality and machine roboetike need right now.
But it is different: Stephen Hawking and Elon Musk signed an appeal to an artificial intelligence paid attention to roboetika and machine morality ceased to be marginalized in the public consciousness of people to think about the safety of technologies and focused on the real problem. Global warming, food and energy crises, over-population and an aging population, the widespread stagnation and a sharp rise in unemployment - all this is much closer to reality than intelligent machines, who decided to destroy their creators. Artificial intelligence will help to solve them.
The best minds are not silent, afraid to go for the paranoid, and openly say they think about the laws governing artificial intelligence, as well as morality and machine roboetike need right now
IN ADDITION, MACHINE RESEARCH ETHICS helps people to reflect on their ethics, and to seek to eliminate the contradictions in it. We'll have a better understanding of how to operate the empathy to understand whether we can trust the robot professions such as nurses, doctors, judges and police. While good arguments on both sides. Although empathy - a mandatory feature for all these professions, the same robots judge could be truly impartial and eliminate discrimination, in some cases, in the long term it is a new institution such as trial by jury. And robots that monitor the elderly in nursing homes, are already experiencing, and people can easily build an emotional connection with them. Be that as it may, the time correction is: all technologies are introduced one by one.
In addition to the confrontation of robots and humans have a far more attractive scenario - such as peaceful coexistence. Before come futurologists predicted technological singularity, people can figure out how to connect to the artificial intelligence directly aligning their benefits with the benefits of the machines. American microbiologist Joan Slonczewski, in turn, suggests that people could go the way of the mitochondria, which were once separate bacteria, but in the course of evolution have become power plants of cells in more complex organisms. In this situation, you can look different, but human evolution will continue, and recent research scientists show that the evolution restart we it from the beginning, most likely would have developed in the same way and in the world today would have lived about the same most mammals. Encouraging aspect is that the best minds are not silent about it, for fear of get paranoid, and openly say they think about the laws governing artificial intelligence, as well as morality and machine roboetike need right now.
