A recent report prepared by the Department of Ethics and Emerging Technology of the California State Polytechnic University for the Office of Naval Research of the Navy of the United States, warns about the dangers of the use of robots in combat and makes suggestions on the code conduct that should have, which could not be easier with the growing advancement of technologies in artificial intelligence. According to this study, one in the cognitive answered is how the algorithms used in real cases, where the consequences of a mistake can be catastrophic.
So I open this paragraph with the open after a controversial incident in Iraq with American soldiers and military robots, which has gone unnoticed by most media.
The U.S. Army has decided to withdraw combat robots deployed in Iraq, after an incident in which one of these devices began to move his weapon without instructions for the maneuver.
The robot is involved in the incident SWORD, a remotely operated robot that can be equipped with a wide range of weapons, although the preference is a 50 caliber machine gun.
These robots are driven by radio, and the incident was that one of them began to move his weapon without the operator for operating the controls. The most likely cause of failure is a software problem.
Although the robot did not open fire, and consequently there was no risk to the soldiers, the possibility that a device equipped with a machine gun can get out is worrying enough for the military has decided not to continue using these robots.
Iraq had deployed four of these armed robots, but planned to deploy 18 more.
The SWORD is a modified version of Talon, a robot that support the U.S. military has been using since 2000 in surveillance missions and deactivation of explosives.
"The military autonomous robots that will fight the wars of the future must be programmed to follow a strict" military code. "Otherwise, you risk the world to suffer untold atrocities at the hands of their steel."
A report prepared and funded by the Office of Naval Research, a secret high-tech and owned by the United States Navy, makes this stark warning and also includes considerations about a possible scenario in the style of 'Terminator' in which robots turned against their human owners.
The report, the first serious work of this nature on the ethics of robot soldiers, to foreshadow a future increasingly close, an era in which robots will be smart enough to make decisions on the battlefield so far reserved for humans. In the end, he warned, the robots could develop significant cognitive benefits on the soldiers of the species Homo Sapiens.
"There is a very common mistake is to believe that robots will do just that are programmed to do," says Patrick Lin, the coordinator of the report. "Unfortunately, this belief is seriously out of date: data from a time that the programs could only be written and understood by a single person."
The reality, Lin said, is that modern programs have millions of lines of code to write and teams of programmers, none of whom knows the program. Consequently, no individual could accurately predict how they would interact with different parts of large programs without extensive testing on the ground, an option that could either not be available or may be deliberately disabled by the designers of the robot soldiers.
The solution, he suggests, is to combine a standards-based programming with a period of 'learning' of things that can and can not be made in the war.
The report covers a wide variety of scenarios in which ethical issues looming, legal, social and political issues that will arise when the robot technology progresses. How do we protect our armies of robots against hackers or terrorist gangs of software bugs? Who gets blamed if a robot were crazy before a crowd of civilians: the robot, its developer, the U.S. president? Should robots have a 'suicide option' or should be programmed to defend your life?
The report, coordinated by the Department of Ethics and Emerging Technologies for Politéctica California State University, the U.S. military harshly warns against complacency and shortcuts, now that the designers of military robots are involved in the race to reach the market and is increasing the pace of advances in artificial intelligence.
The sense of urgency among the designers may have been accentuated by the Congressional mandate in 2010 to one third of all aircraft are operating drones attack in 2015 and they are also one third of all infantry fighting vehicles .
"The race to reach the market increases the risk of an inappropriate design or programming. Worse, without a significant and sustained effort to instill ethical codes autonomous systems, there is little hope that the first generations of such systems and robots are appropriate, therefore commit mistakes that could cost human lives, "warns the report.
A simple code of ethics in line with the 'Three Laws of Robotics "which ran in 1950, Isaac Asimov, the science fiction writer, will be insufficient to ensure ethical behavior by the autonomous military machines.
"We're going to need a code," says Lin. "These gadgets are military and can not be peaceful, so we have to think in terms of the ethics of war. We're going to need a warrior's code."
Principles or laws of Robotics, Isaac Asimov set in his fictional novel, written in 1950, "I Robot" were:
1) A robot must not attack a human or with their inaction, allow a human damage.
2) A robot must obeceder orders given by human beings except where such term is in conflict with the first made up of laws.
3) A robot must protect its own existence as long as it does not conflict with the first and second of these laws.
However with the gradual integration of robot soldiers in armed conflict breaks the first law in the larobótica fiction novel by Asimov, which have already been incorporated in the case of the Korean Government.
THE ROBOT PROTAGONIST MILITARY SWORD OF INCIDENT IN IRAQ WITH U.S. soldiers.
INCIDENT IN IRAQ WITH MILITARY ROBOTS WITH HUMAN
Three Laws of Robotics
In science fiction the Three Laws of Robotics are a set of rules written by Isaac Asimov, which most of the robots from their novels and stories are designed to meet. In this universe, the laws are mathematical formulas printed on the trails positronic brain "of the robots (which today would call ROM). First appears in the story Runaround (1942), provides:
A robot must not injure a human being or, by inaction, allow a human being suffers harm.
A robot must obey orders given it by human beings except where such orders are in conflict with the First Law
A robot must protect its own existence, so far as such protection does not conflict with the First or Second Law
This formulation of laws is the conventional way in which the human stories of the state; would be the real form of an equivalent set of instructions and more complex in the brain of the robot.
Asimov attributed to the three Acts John W. Campbell, who would have drafted during a conversation held on December 23 of 1940. However, Campbell argues that Asimov and had thought, and simply expressed between the two in a more formal.
The three laws in a number of stories from Asimov, as shown in its full range of robots, as well as several related stories, and the series of novels out by Lucky Starr. They have also been used by other authors when they have worked in the fictional universe of Asimov, and frequent references to them in other works, both fiction and other genres.
Purpose :
These laws arise as a measure of protection for human beings. According to the Asimov, the conception of the laws of robotics want to counter an alleged "Frankenstein complex," ie, a fear that human beings developed compared to the machines that hypothetically might rebel and rise against their creators. Even trying to disobey a law, the robot's positronic brain be irreversibly damaged and die robot. A first level presents no problem to equip robots with such laws, after all, are machines created by man for his service. The complexity is that the robot can distinguish which cover all situations that the three laws, that is able to deduct at the time. For example in a particular situation to know whether a person is running or no damage, and deduct what the source of the damage.
The three laws of robotics represents the moral code of the robot. A robot will always act under the imperatives of his three laws. For all intents and purposes, a robot will behave as a being morally correct. However, it is legitimate to ask: Is it possible that a robot violates any of the three laws? Can a robot "damage" to a human being? Most of the Asimov robot stories were based on situations where despite the three laws, we could answer the above questions with "yes."
Asimov created a universe in which robots are key to over ten thousand years of human history and continue to have a decisive role for ten thousand years. It is logical to think that the level of development of robots vary over time, increasing their level of increasing complexity.
History of the Three Laws of Robotics
The first robots built on Earth (viewed, for example, I, Robot) models were developed recently. It was a time where robopsicología not yet developed. These robots could be faced with situations in which they were in conflict with their laws. One of the easiest situation is when a robot to harm a human being to keep two or suffer more damage. Here the robots decided on the basis of a purely quantitative approach, then left unused, to be forced to violate the first law.
Subsequent developments in robotics, allowed the construction of more complex circuits, with a greater capacity for self-reflection. One peculiarity is that robots could redefine their concept of "harm" as their experiences and to determine levels of it. His valuation of human beings can also be determined by the environment. So a robot could injure a human being to another to protect it considers most valuable, particularly his master. It could also be the case that a robot to physically harm a human being to another to avoid being damaged psychologically because it becomes a tendency to consider psychological harm more serious than physical. These situations would never have occurred in older robots. Poses Asimov robot stories in the most diverse situations, always considering the logical possibilities which could lead to robots with such situations.
The Act Zero
One can reach with fond Asimov robots, which shows in their stories robots increasingly "human". In Bicentennial Man, Asimov tells the story of Andrew Martin, born robot, and that struggle throughout his life to be recognized as a human being. Also R. Daneel Olivaw and R. Giskard Reventlov, which have a key role in the second expansion of humans and the subsequent founding of the Galactic Empire. Being the most complex robots ever created, were able to develop the law of robotics zero ( "Zero-law" in English) as a corollary of the first philosophical:
A robot may not harm humanity, or, by inaction, allow humanity to suffer damage.
R. Giskard died in Robots and Empire, after being forced to harm a human being under the law zero. The fundamental problem with this law is to define "humanity" and to determine what constitutes a "harm" for Humanity. R. Daneel achieved through the sacrifice of assimilation Giskard, since becoming the protector in the shadow of humanity. Daneel under different identities, he became one of the most important cycle Trantor (formed by the robot stories and novels, novels of empire, and the saga of foundations: 17 books) is also a key element in continuity.
Criticism
In principle it should be considered for autonomous intelligent machines that violate the law.
A missile can be considered intelligent and has no problem killing humans. A computer system that coordinates air strikes and land use in a military operations center violates the law. A police robot that could be created should be able to kill like the police man. This would involve reviewing the fundamental concepts that underpin a society. For example, if private property premium on human life, then the police should be able to kill humans to protect private property. That is a robot or a human did not make a difference because the result is the same.
The problem arises when considering a set of robots can take over the resources and enslave or annihilate the human no matter what humans are concerned. Any human regardless of race, religion or economic status would be treated the same way. This leaves aside the problem of humans who enslave one another and the genocide being committed against certain groups. Would it be more serious than what makes a group of robots?
In general science fiction movies like I, Robot pose an extremely intelligent robot may decide to control the destiny of humans and that this is unacceptable. These films suggest that the thought of a completely logical robot is unacceptable because emotions may regulate the conduct of a most appropriate way, more human.
It should be stressed that this would contravene the natural evolution. If robots are human creations and beyond us in intelligence and ability, then that humans are annihilated not pose a problem, because the robots would be a higher evolutionary state, and humans have been the missing link between apes and robots. The robots are evolving field.