The Pentagon can not do without artificial intelligence

The Pentagon has launched a large-scale initiative related to the development of systems that will extract useful experience "from large and complex arrays of digital data." The amount of the initial investment will be 200 million dollars and will be sent to 6 various government agencies. The new project will be an addition to the XDATA program of the defense science agency DARPA, the cost of which is 60 million dollars a year. This program is aimed at automated intellectual analysis of large amounts of unstructured data that are of a different nature (for example, road traffic data), cnews.ru reports.

The US military has long been faced with the problem of a huge data stream, which analysts have to disassemble. A large amount of information that comes from satellites, from intelligence, from unmanned reconnaissance planes, from the Internet by dozens of terabytes leans on the analyst. This dataset contains a lot of useless information, but because of the huge amount it is possible to miss very important information.

At the same time, there is a rapid increase in the volume and complexity of information. Today we have to study the fragments of documents, hard drives, aerospace imaging, radio interception data, etc. The data that is collected is often imperfect, incomplete and heterogeneous. The aggravation of this negative trend is due to the proliferation of the Internet and various digital devices. In such a volume of data, terrorists can easily hide their plans and communicate with each other.
In this regard, the Pentagon intends to rely on a new generation of computer systems that will be able to interpret the real world with the “wits” of a person and with computer speed. At the same time, the US military is going to refuse to use only powerful centralized analytical systems; they plan to endow various platforms, such as unmanned aerial vehicles, for example, with limited intellectual capabilities.

First of all, these are "visual intelligence" systems on military aircraft. The use of special information processing algorithms will make it possible to leave unimportant events aside and notify about important events. Computers endowed with elements of artificial intelligence will be able to independently mark important information through appropriate comments, and, most importantly, they will automatically detect and recognize the enemy. Long-term large-scale use of such systems in the future will make it possible to "instruct" robots not only to monitor the enemy, but also to destroy him automatically.

Comments

[Dmitriy69]

The lack of own intelligence has to be made up for by artificial

[marauder]

Sawing money again. Then they say the program failed, the money ran out.

[755962]

Quote: Maroder

Sawing money again

Strange as it may seem, if you dig deeper, the Morgans and the Pentagon are a single whole. And if we talk about traditions, the closest connection with the military business with its financing is perhaps the most important and persistent tradition in the activities of the Morgan family. For a short period of about 10-12 years, several managers have been replaced at the US Department of Defense, the body that runs the arms race and distributes military orders between companies and firms. But by a strange “coincidence”, each time the head of the ministry is either a direct representative of Morgan companies or a person close to them. The first Minister of Defense in the Eisenhower government was Charles Wilson, president of the Morgan-Dupont General Motors Corporation, the same Wilson who glorified himself with a cynical formula: “What is good for General Motors is good for the country.” In his mouth, this formula with the same reason could have sounded like this: "What is good for the Morgan is good for the United States."

[virm]

It's actually great. It seems to me that it is artificial intelligence that will bring a transition to a new technological level. Not nano, not bio, not alter energy. Moreover, the creation of AI, or at least AI elements, is a theoretical and algorithmic task, and not a technological one at all. To date, there is a sufficient level of technology to implement the elements of AI. Only a good (very) idea and large-scale implementation at the software level is required.

[Dimitrxnumx]

AI will be invented in the Pentagon, they will call it skynet, and there robotic platforms will approach, skynet will take control of nuclear weapons and begin .... somewhere I already heard it, or rather watched) ...

[vylvyn]

It’s interesting, buddy, but you will like it when this AI, having moved to a new technological level, scans you and all your close ones and relatives, sees internal organs that are not quite healthy thanks to the ecology, and decides to send everyone to the cemetery so that you don’t spoil the environment further. Then I'll see how you say it - Actually it's great.

Once upon a time, people came up with biological, bacteriological, chemical weapons. Thank God that we agreed not to use it, but to destroy it. For AI, too, there must be a limit that cannot be exceeded, otherwise we get tired of running away from terminators, matrices and drones. And no Jedi will help.

[Opertak]

Quote: virm

Today there is a sufficient level of technology to implement the elements of AI

To date, there are no prerequisites for the creation of AI. Who says otherwise - and the ignoramus. I will say more, in the last 12-15 years, computer technology is almost at a standstill and in the near future its full stagnation will come. The creation of multiprocessor computing systems is only a sign of this, an attempt to delay the full kirdyk and nothing more.

[Dimitrxnumx]

In general, if it turns out to invent AI in the understanding that we represent, then there will be a lot of applications for it, not only in considering and sorting incoming information. Then there will really be a "breakthrough" in science and, later, technology. But the military will pervert everything ...

[Arkan]

Chinese hackers will be happy.

[Anubis]

Over 50 years of development in the field of artificial intelligence have not advanced a single step. I’m not saying that they didn’t give anything - speech recognition systems, images, texts and much more. But it is AI - the prospect is as far away and unrealistic as it was 50 years ago.

[Opertak]

I agree to all 1000% !!!

[Neolexx1]

I do not want to offend anyone. but what do some wise men of general know about the principles and degree of progress in the development of AI by various organizations on our planet? Do you even know why multicore systems are being developed? Parallelization is the point of everything! Significant increase in speed and time efficiency with less energy consumption. And as for the article, I’ll just say that the pictures from Terminator 4 until 2030 may not seem so fantastic ...

[Anubis]

No one is offended. And about the progress in the development of AI - read here http://www.forum.aiportal.ru/forum4.html in short:
- for 50 years, new approaches to the creation of AI have not appeared, despite the fact that the old ones have shown their inefficiency (speed does not solve anything here, if everything rested on the speed of calculations, AI, albeit slow, would have been created for a long time)
- with all the progress in the study of the human brain, so far no one can say where the area responsible for thinking is located in the brain, not to mention how this thinking works
The creation of AI turns into an attempt to create a device whose operating principle is not clear either in the details, but in general.

[Honory]

Given the current level of AI, Americans are happy early.