"One programmer can control an entire army," explains Prof. Toby Walsh, director of a research group at wnsw, the University of New South Wales. "All weapons of mass destruction are already prohibited in development, or in the process of such prohibition - atomic, chemical and biological weapons. We must add the autonomous weapons to this list, which we are morally unable to accept."
Leading researchers in the field of robotics and artificial intelligence from two prominent countries in the British Commonwealth - Canada and Australia, recently signed open letters calling on the heads of government to take a stand against turning artificial intelligence into a weapon.
In an interview with the Hidan site, Prof. Toby Walsh, director of a research group at wnsw, University of New South Wales, explains that the letter was prepared in preparation for the UN Disarmament Conference held in Geneva in November 2017.
So far, 19 countries have signed a ban on the development of autonomous weapons; Algeria, Argentina, Bolivia, Chile, Costa Rica, Cuba, Ecuador, Egypt, Ghana, Guatemala, Vatican, Mexico, Nicaragua, Pakistan, Panama, Peru, Venezuela, Zimbabwe and the Palestinian Authority.
Autonomous weapons are also described as killer robots, to burn a third description into the minds of most humans. "We are not talking about an exterminator in the cinematic style of 'Terminator'," explains Prof. Walsh, "but much simpler technologies that can be implemented within a few years. For example, a predatory drone that will fly over the skies of Iraq without human remote guidance but by a computer. Now, in fact, a computer will be able to make the final life and death decision - whether to fire the missile - or not."
According to him, "I am not afraid of a smart artificial intelligence, but rather a stupid artificial intelligence." We will give machines the right to make life and death decisions, but the current technology is not capable of making correct decisions."
"In the long run, autonomous weapons will gain greater capabilities - but I am concerned that someone will turn them into weapons that will undermine the geopolitical order, and eventually become another type of weapon of mass destruction. The Australian letter was distributed simultaneously with the Canadian letter addressed to Prime Minister Justin Trudeau, which was signed by two of the founders of the field of deep learning and artificial intelligence - Jeffrey Hinton and Joshua Bengio.
Third military revolution
"I organized the Australian letter that was signed by 122 scientists, including dozens of deans and heads of university departments, as well as professors of AI and robotics," Walsh said.
In the letters, the researchers write that a deadly autonomous weapon that lacks human control is on the wrong side of morality. According to Prof. Walsh, they appealed to their heads of state to commit to working with other countries to include autonomous weapons as one of the prohibited developments, similar to biological weapons and chemical weapons.
He says that "the UN is listening to these voices, which have been heard for several years, and it is responding." But like all things in diplomacy, development is not fast."
In December 2016, after three years of informal talks, the UN began a series of formal talks within a group of government experts that included technologists, law and political experts selected by member states to recommend action, but not to negotiate an autonomous weapons non-proliferation agreement.
The danger in the non-existence of such an agreement, says Prof. Walsh, "is that an arms race will develop to develop more and more powerful autonomous weapons. This will be a third military revolution - after the invention of gunpowder and the development of the atomic bomb. This third revolution will be another step in the speed and efficiency with which such a weapon can kill."
"One programmer can control an entire army", he explains. "All weapons of mass destruction are already prohibited in development, or in the process of such prohibition - atomic, chemical and biological weapons. We must add the autonomous weapons to this list, which we are morally unable to accept."
Prof. Toby Walsh, Research Group Director at wnsw, University of New South Wales. Photo: Public Relations
Who develops such a weapon?
The United States, China, Russia, Great Britain and other advanced countries, including Israel.
The "raw material" of an AI bomb are talented people. Could terrorist organizations pay them to develop such a weapon? How can we audit this?
"The way to criticize these weapons is the same way we criticize other technologies such as chemical weapons and biological weapons. We could not stop promoting chemistry, but we made the use of chemical weapons morally unacceptable, and weapons manufacturers were prohibited from selling them. This way they are not for sale nor available on the black market, so their use is extremely limited. We can hope that the same thing will happen with the autonomous weapon as well."
"We cannot stop the development of artificial intelligence. It can be used for many peaceful purposes such as autonomous cars, but we can make it immoral and unacceptable to use it to kill, as we did with chemical and biological weapons. In this way, I hope, we will make the world safer."
What do you know about what is happening in Israel in this field?
"Israel will not be safe if and when autonomous weapons exist. They will have to defend themselves against this terrible weapon."
The article was published on the People and Computers website as part of a weekly section dealing with artificial intelligence.
14 תגובות
Rival, IV
There are many weapons that are prohibited for use, and those who break these laws can find themselves in serious trouble, so the laws have meaning.
On the other hand - autonomous weapons have always existed, since ancient times. Don't we prefer mines that neutralize themselves at the end of the war? Or aircraft missiles that will know not to shoot down such friendly aircraft (and this has been around for a long time).
Such letters are important, to bring up the problems. Is it better to ignore the dangers? It should be understood that "artificial intelligence" has no definition, so it really doesn't have much practical meaning.
Every operating system today has artificial intelligence, every phone and every car.
anonymous
Atomic weapons were invented in the US because they were afraid that someone else would invent them first. It is an offensive weapon designed to do maximum damage, both to property and human life.
There is no evidence that nuclear deterrence saves lives. Whoever claims otherwise - let him prove it.
This is a worthless document that has no relevance in reality. Artificial intelligence is a tremendous force multiplier that no army and no army commander would be willing to give up on the modern battlefield. Already today there are quite a few elements of artificial intelligence on the battlefield, for example the Israeli Hagil missile (from the Spike family) uses artificial intelligence to hit targets both in difficult visibility conditions and when the target tries to hide or evade, the F-35 plane we received has quite a bit of artificial intelligence that is without it The plane is worthless.
Who would be willing to give it all up?
122 scientists signed a document…
Even if 122,000 scientists sign this document, nothing will change. The work of those scientists themselves will be used to build weapons of mass destruction or some other despicable and stupid plan, because the decision makers will always be greedy and short-sighted people. (Proverbs of Yotam, Book of Judges, Chapter XNUMX)
It is quite clear that the Palestinians have no problem opposing nuclear weapons.
After all, they don't have him.
But the State of Israel has and has. And also to the USA.
In general,
No problem with nuclear weapons.
As there is no problem with a car.
And just as there is no problem with a knife.
The atomic weapon was invented to protect the Allies, against the wretched of life (at the time, Hitler would destroy his name and today, with Abu Adolf Mazen, his house will be destroyed on him and his family)
The problem is when these things fall into the hands of wretched souls. Like the Palestinians and the rest of the world's terrorists.
I have a problem with this article...
It seems that as time goes by, indiscriminate mass killing does not belong in the world we live in.
Man today is the subject of occupation.
The occupation is mainly financial.
And since that's how - the elimination of a person is similar to the liberation of territory (which years before had been fought for)
Man today - throughout his life must use the services provided by the corporations (whether it is medicine, nursing, technology, etc.). And since that's how a dead person is worth almost nothing.
It follows that as time passes, people will live longer (5-year-olds today will undoubtedly live more than 100 years) and the capabilities of the corporations will only get stronger.
The question arises - what is the role of regulation (the legislature) in this whole equation?
These should look like they are facing a citizen
But in practice (and yes, there should be an emphasis on the p.) these serve the corporations.
And this whole story can be called ML-3 (World War III).
----
The problem with this war is: that we know how to recognize a war only when there are tanks, planes and other destroyers and in the war I'm talking about these have no place.
To kill a person is to burn money - and money is the thing for which the war takes place.
———–
{I might be talking nonsense... but that's how I feel. This whole section of assassinations may be related to terrorist organizations, but it is no longer really related to the world we live in}
Asimov already realized over sixty years ago the futility of limiting "dangerous" technologies - most of the time they are simply bypassed and reach the same exits in a much cheaper and simpler way. Something these researchers are simply unable to grasp.
So in 20 years or so, Yossi will have no problem asking his artificial intelligence to download a model of a drone, print it on the home XNUMXD printer and order it to drop plastic bags full of water right on the head of his annoying neighbor David who insulted him on the neighborhood WhatsApp. Now try to find an artificial intelligence that can prevent such cases, or more serious ones.
Apart from that, I saw that the Palestinian Authority signed this treaty. This is no wonder considering the fact that they have no problem recruiting people with natural intelligence to serve as expendable actuators for their explosive belts and inferno trucks.
Nostradamus
You are far from being Bill Gates... you are a person who lacks understanding and eats hatred. Because of fools like you, violent Islam is so successful.
The whole enlightened world is against you... my poor...
All these restrictions have one purpose. To limit the power of the West to fight the barbaric and stupid Islamic Satan who is unable to develop such technologies
Don't confuse the brain, you are as right-wing as I am Bill Gates. Like peace now they want peace, like the left is Zionist and like the Turkish military operation is called an olive leaf. This is exactly the leftist tactic, to use words of peace and make war to confuse the enemy. The Muslim countries sign the Hudayba agreements, false agreements like Muhammad taught them. In the Oslo Accords Arafat promised to fight terrorism and the next day he armed everyone. They appoint and establish terrorist organizations that have no problem using any possible weapon
Nostradamus
Are you embarrassing the right again? Did you notice that a large part of the countries that signed the treaty against autonomous weapons are Muslim? And the list of countries that are not signatories to the anti-landmine treaty includes Israel and the USA?
Autonomous weapons have been around for hundreds of years. Already in the thirteenth century the Chinese used mines. And a thousand years before Gan there were already traps such as nets and covered pits.
And even then - it was a terrible weapon.
Of course, because Muslims are allowed all types of weapons and the West is not allowed to win, we need to create a balance of power. Because the Muslims are weak and oppressed.. The leftist discussion is distorted and abominable
One thing is certain that no one in the signed list of countries will develop artificial intelligence (and the PA is not a country),
But beyond that, I doubt if there is a mechanism that can be built and stop it because there is no digital transition from 0 artificial intelligence to 1 artificial intelligence like in the development of nuclear weapons where either there is or there is no weapon. In addition, we see what the ability to control the development of nuclear weapons is worth, in artificial intelligence The transition is more of a developmental analog without a clear limit, something that will be much more difficult to supervise, how can a system be classified at all, has it crossed the threshold or not
or has the potential to cross the threshold,
For example, a defensive system can also be used as an offensive system,
For example, banning an artificial intelligence system that can protect against an unconventional missile barrage will be banned and thus hundreds of thousands of people will die to satisfy the experts' sense of justice and morality? After all, it is clear that such systems will not be built in a big way
And they also have an offensive projection built into them, which may not be used in a low-intensity conflict, but certainly in a conflict
General will be activated, apart from that anyone who was a corps would not hesitate even for a tenth of a second whether to bring robots into the battlefield or to sacrifice an animal on the altar of the experts' justice, on a tactical level it would be immoral to send a corps at risk if the same operation could be performed with a robot, which is certainly not of interest most of the experts because it is likely that the most dangerous thing they have seen in their life is the computer keyboard, there is also a slight suspicion that some of the experts have
Also an anti-Western tendency that dictates this campaign and not only the fear of the rise of AI-based combat systems,
It would not be surprising if they clothed Israel and its defense systems in the guise of "sincere" concern for the human race.
Despite all of this, I would not dismiss everything they say. We definitely need to hear different opinions because there are risks that even without being an expert you can understand them, but I am skeptical if it will even be possible to build some kind of effective model that is able to monitor such systems also because of the international political aspects.