War or peace - computers decide

December 2 2010 1

Declaration of war or peace - the solution of such crucial issues is going to be entrusted to machines. In the UK, they are developing a computer system, which is essentially the same as that known to all fans of fiction on the blockbuster Terminator. As you know, in the film, humanity designed the Skynet defense supercomputer with artificial intelligence, which once decided to unleash a global war against people.

The military intelligence machine currently being designed is already named. The system, which can theoretically manage combat operations and make global decisions about war and peace on Earth, is called ALADDIN (an abbreviation for “Autonomous Learning Agent for decentralized databases and information networks”). It is curious that in terms of its technical characteristics it will have to almost completely coincide with the supercomputer that appeared in the Terminator.

The system will be a decentralized network of computers equipped with a self-learning intelligent agent-program, which independently performs the task specified by the user for long periods of time. Intellectual agents are used in computer science, in particular, for the constant search and collection of necessary information.

Actually, strategic computers themselves have long existed. Suffice it to recall chess machines or "intelligence" in strategic video games. But now scientists are introducing all these ideas into the defense industry at the most global level.

The system is developed by the British company BAE systems. According to the experts who create ALADDIN, people are no longer able to make adequate decisions in the framework of waging modern wars. The human brain is not able to process too much information that provides it with the latest technical tools from the battlefield. Moreover, these colossal data sets have to be processed as quickly as possible, which puts a person in a dead end. Because of this, the warlords decide to force the river along or open fire on their own. There are quite a few cases when people make the wrong decisions, confusing similar terrain names and the like. According to the British, ALADDIN will not allow such a miscalculation. In addition, the computer will operate with a variety of information, it will be able to calculate the many possible options for combat and choose the best.

Finally, the ALADDIN system would be much faster in contact with robots on the battlefield than a human general. By order of the US Navy, a $200 Octavia robot has already been designed. A machine with a human-like face will have to take an active part in hostilities in the future. For example, she will be able to participate in counter-terrorism operations in Afghanistan as a sapper to clear roads and buildings.

As Alexey Bakuradze, a research scientist at the Information Technology Department of Robotics Management at the St. Petersburg Institute of Informatics and Automation of the Russian Academy of Sciences, told RBC Daily, ALADDIN is likely to be based on multilayer perceptrons (a cybernetic model of the brain): “There is an algorithm for recurrent learning in real time, which allows system to quickly learn and effectively solve problems such as command and control, clearly understand the system "friend or foe".

The developers believe that ALADDIN will be successful not only for strategic planning of the operation, but also for making global decisions from the category of “to fight - not to fight”. And here it remains unclear - in a situation where people could agree whether the cold-blooded system would like, having assessed all the parameters, give the command: “Fire!”