Combat drone X-47B. Image: Dod
According to a report on the worldwide development of autonomous systems for the militar, France and Germany are also advanced in addition to the US, Russia and China
From the Pentagon hoard the insurance that one increasingly sat on autonomous systems, but they should not be really autonomous. Never, said recently Defense Minister Ash Carter will use the Pentagon fight robot without "Man in the loop" Select whether to fire or not. And also in the "Third Offset Strategy" It is emphasized, autonomous systems came into question mainly in the cyber area, otherwise it is primarily about the development and improvement of the interfaces (US Secretary of Defense: "Never" Full autonomy for combat robot).
Such an interface is approximately the plan to accompany manned F-35 combat aircraft of unmanned F-16, which can be controlled by the pilot to ward off attacks or can attack more effectively. For the time being, tempered F-16 was rebuilt to unmanned aircraft and serve as target objects for training.
Probably the reasons for rearing the public and its own troops that do not want to drive autonomy and want to replace soldiers by robots, but is questioned by a report. The Pentagon ahead of the development of artificial intelligence, in order not only people as actors, but increasingly to replace human decision-making.
The report was funded by the Future of Life Institute, which is funded by Elon Musk, the land of the Space company Spacex and the Eauto manufacturer Tesla, who wants to realize a manned ride to Mars, but warns of autonomous systems and with the Founded Research Institution Openai, which is founded in 2015 the question of "Existential threat of artificial intelligence" and "Safe Ki" want to make.
The report was written by Heather Roff (Global Security Initiative ASU) and Richard Moyes, Director of British Organization Article 36. It is the attempt to make an inventory over the development of autonomous weapons systems worldwide, using all possible sources. The main actors are according to the report in the order: the USA, Russia, China, France and Germany. Created and described are 284 systems that can already act partly autonomously.
Most autonomous combat systems are still missiles
Under a full autonomy, the authors will understand systems that can be moved independently in the environment to any location, selecting the targets in their environment and will fire them, and they can produce and generate their goals, including the observation of their environment and communication with other agents. A "autonomous" System has been many of the properties have to have goods, however, far away from owning a human intelligence. The systems are usually missiles, the most important capacity is the launch control. Here is a difference:
- The system can not push himself, this must make a person or a person is only informed;
- The system informs a person and waits for the command to start firing, or man eliminates the goal and fires;
- The system is set so that it fires on a target when a person does not stop it;
- The system slides without human action.
Most of the existing systems existing or developed in development can, however "only" After shooting find the way to your intended goals. Although it has been there for several decades, but the outcome of many newer KI developments is. In addition to self-mobility, the authors still differentiate self-direction, ie the detection and selection of goals, as well as self-determination, D.H. The ability to set goals and urge what some possibility involves having to make decisions.
Drones like Neuron (EU), X-47B (USA) or Taranis (UK), which float over an area (Liter), compare objects on the ground with a database and highlight a destination if a match has been found, have been found under self Direction fall. This is referred to as the new front of autonomy, as not a given target is searched here, but searched for many goals and how a Jagery is waiting for one of them to be discovered. So far, this is only installed in a few weapons systems, but the technology is a strongly explored area.
Above all, it is currently being sought to improve the possibilities of target identification, to distinguish friend and enemy and learn to demand: "Systems have to be able to adapt, learn and plan or update. In short: The systems must have more things and inaccurate tasks exports."
Stealth Property is the primary motivation to develop autonomous systems
Autonomy is currently not developed, "To fight in addition to people on the battlefield", Say Heather Roff, "but to replace them. This trend becomes stronger, especially for drones when studying weapons in development." In drones, for the trend towards autonomy, an essential role is played that they can also be used against Lander, which can recognize and enhance drones highly flying with air defense systems. It is therefore important to allow the property to fly undetected from the opposing radar and without rendering control commands, but it should be able to communicate with other systems as needed and to uplift targets.
According to ROFF, the Stealth property is present at present a primary motivation to develop autonomous systems. The coarse states do not have to have the nose in front with most resources such as the US, Russia or China. If regional factors drive the use of such systems, this is not exactly stabilizing, my authors. "Regional power with coarse capitals in the development of autonomous weapon systems like Israel", so, "can destabilize a region through their use or export to other states."
But how Patrick Tucker writes in Defenseone, the destabilizing effect comes to advantage, especially when the one side uses autonomous systems that can make quick decisions because no one is more involved in the decision-making process. Not to join in here, be morally attending, but konne connect in a strategic disadvantage when an opponent does not scare about it. Tucker points to the high-frequency trading in which people are already before, because too slow, already.