Tricky comments. Where to push artificial intelligence in the absence of ordinary?
Of course, the whole report does not make sense to read. It is huge, but, as a matter of fact, no one in the United States really reads it and will not, for this purpose just intelligence is needed. At least in order to understand where the truth is, and where is nonsense. Enough short husks.
In general, we are talking about the fact that the United States nevertheless achieved some success in repelling cyber attacks directed against government structures and services. And soon the thesis of the novice hacker on the topic "How to hack the Pentagon server" will really be the lot of the chosen ones.
However, based on their success in the fight against hackers, the speakers (C. Fields, R. Devil and P. Nielsen) say that in the coming 25 years, the United States is simply obliged to allocate additional allocations for AI development programs.
Why?
"... the country needs to take immediate measures to accelerate the development of military technology of artificial intelligence. Academic and private studies of artificial intelligence and autonomous technologies are far ahead of the development of the American armed forces. There is a real chance that we will see a repetition of what happened in cyberwar, where the US was focused on repelling the attacks, and the struggle was not easy, as we were taken by surprise. "
The argument of the intellect does not shine, if so, seriously. What, excuse me, AI, if the usual hackers could barely cope? However, the report was heard and commented on in the media by Robert Wark, whom we had not once sent. He gave his comments to the "Financial Times".
Wark, commenting on the speakers' arguments, says that the Pentagon needs to collect more analytical information about the AI capabilities of other nations, as well as develop solutions for "reciprocal autonomy." In addition, the authors suggest that the United States is simply obliged to contribute much more resources to the development and testing of AI training, such as autonomous weaponthat adapts to battle conditions.
Hi, Skynet? Someone "Terminator" seen enough? So it would be worth watching the whole serial. There it was well shown where they brought games with AI in general and SkyNet in particular.
But the conclusion is unequivocal: it is possible for the US hackers to resist, but, God forbid, the enemy AI - no chance. And, accordingly, "you need more gold!"
And from where, forgive, this enemy AI will suddenly be drawn? It's simple for authors.
As an argument by the gentlemen, the speakers cite as an argument that competitors, namely Russia and China (attention !!!) do not have a policy of restrictions in the development of independent lethal military units.
Boom! No, it's so clear where all the world's evil for the United States threatens with a cataclysm, no problem. There is only one conclusion: if it is not forbidden in Russia and China to create robots-killers and endow them with artificial intelligence, then, of course, they will do it.
Otherwise, why haven’t these Russians and Chinese imposed 10 thousands of bans on the development of AI? Similarly, 146% is the minimum, that in the secret laboratories of Russia and China hundreds of scientists are already sitting and puzzled over the projects of robots endowed with AI and capable of destroying all life.
"This does not mean that the US should follow their example, but it may need technology that can prevent deadly AI systems before it is too late."
No, well, I applaud the authors. There is no information about the presence of any AI in Russia and China. There is no information about killer robots and other figatory either. We are also silent about the data on combat robots with AI, but we need to write down the system of countering these robots today. Tomorrow may be late.
The most convenient arguments for requesting funding for long-playing projects. Really, what's wrong with getting yourself a job for the next 25 years? Given that you do not need to do anything, then I think you can get along pretty well.
As scientific advisors to the US Department of Defense suggest, the key areas for the development of military equipment for the next 25 years are artificial intelligence and autonomy.
According to Wark, the most promising area of research is the increased interaction between artificial intelligence and man, when a decision is made jointly. That is, this same AI and the person will jointly decide for whom to open fire, and for whom it is not worth it. Perspective is still ...
Glory to the Almighty, however, it is not a matter of giving the military robots complete autonomy. According to Wark, the next 10-15 years of the US military intend to focus on the development of machines that can act and make decisions only within certain, predetermined parameters.
Again, according to Wark, the same work is being done today in the military laboratories of Russia and China, but the United States has a number of advantages compared to its competitors, including much more research experience. But - "you need more gold!"
Another question that worries the Pentagon is how to keep its own technological developments secret for as long as possible. Considering what was mentioned at the beginning, the hacker community’s increased attention to the secrets of the United States and the clearly inadequate protection capabilities.
But even here, Wark has an answer that, intellectually, pulls on heels of kilopsak. Wark says that the trick (a trick, yes!) Is to “demonstrate opportunities that are serious enough to make rivals doubt that they can win.” At the same time, "major developments capable of having a decisive influence on the course of hostilities must be kept under the strictest protection, like a military reserve."
And this is the deputy minister of defense ... Artificial intelligence, speak? Normal would get to start ...
After that, we can’t even turn our tongues about cutting. Kindergarten, younger group.
Materials used:
http://www.acq.osd.mil/dsb/reports/DSBSS15.pdf
Information