Robots that beat people at chess or computers that we can talk to – for many years scientists have been trying to artificially recreate the complex human mind. How far did they get with it?
How intelligent is artificial intelligence?
The research field “Artificial Intelligence” (AI) tries to imitate human perception and human actions by machines. What once began as the science of computer programming has evolved more and more into the study of human thought.
Because after decades of research, the impossibility of creating a “thinking” machine has been recognized without first having researched and understood human thinking itself. That is why there are sometimes large overlaps between AI research and neurology or psychology.
To date, it has not even come close to being able to understand human intellectual performance as a whole with machines. A major obstacle is language processing.
Indeed, even the execution of the most straightforward orders is a profoundly mind-boggling process for a machine.
Research is thusly progressively focusing on individual sub-regions, to make work simpler there, in addition to other things. This requires a steady trade between researchers from a wide assortment of disciplines (mental science, brain research, nervous system science, reasoning, and semantics).
Numerous researchers recognize solid and frail AI. Feeble AI covers just incomplete areas of insight. It is for the most part founded on techniques from math and software engineering and is utilized, for instance, in route frameworks, and discourse acknowledgment.
Solid AI includes consistent reasoning, arranging, correspondence, making autonomous, complex choices. Numerous scientists question that this will at any point exists.
Furthermore, regardless of whether it was to succeed, numerous moral inquiries could emerge: Which choices can be passed on to man-made reasoning that has no ethics and no consciousness of right, wrong, and, most importantly, subtleties?
When does a PC finish the Turing assessment?
The subject of when a machine is viewed as keen has been a worry of AI research for a long time. An estimation device that is by and large acknowledged is the alleged Turing test.
It was created in 1950 by the British mathematician Alan Turing: An individual speaks with someone else and a machine in equal over an extensive stretch without visual or hear-able contact – for instance through a talk program.
Man and machine attempt to persuade the analyzer that they are thinking people. If, after the discussion, the analyzer can’t say with a sureness which of the questioners is a human and which is a machine, the machine has breezed through the assessment and can be thought of as keen.
In 1991, the US social scientist Hugh G. Loebner offered a $100,000 prize for the PC program that finished the Turing assessment and tricked a jury of specialists.
In 2014 there were reports that the Russian programming “Eugene Goostman” finished the assessment. Nonetheless, questions about the technique and the test game plan were voiced subsequently.
Starting in 2020, nobody has gotten the honor, and most AI analysts don’t anticipate that that should happen any time soon.
Tamagotchis, Robots, and Co
The areas of use of man-made consciousness are very different. Frequently we are not even mindful of them. Their utilization is best in little sub-regions like medication: robots complete specific careful areas – in the thousandth of a millimeter range – significantly more definitively than a specialist.
In production lines, especially in the automotive industry, robots replace countless human actions. Automated arms, for example, those utilized at General Motors during the 1960s, have become imperative, particularly for assignments that are perilous to wellbeing and inclined to mishaps, like work of art or welding.
Games, especially board games such as checkers and chess, are a classic area of application for artificial intelligence. Programmable and learning toys, mini-robots, and computer programs have long conquered children’s rooms.
The legendary Tamagotchi is already a thing of the past, but other artificial companions such as robot dogs, talking dinosaurs, or dolls are pushing onto the market, with which one can communicate through simple gestures or speech and which carry out specific tasks.
Expert Systems and Machine Learning
Expert systems are specialized in very specific and narrowly limited areas of application. An example of this is programs with which computer tomographic recordings on the computer screen are converted into three-dimensional images. Doctors can get an “image” of the respective part of the body and its condition.
An important component of artificial intelligence is self-learning systems. These are utilized, for instance, in robotized spam channels in the email inbox. You feed a computer with example data, which it evaluates and analyses.
The system recognizes patterns and similarities and can thus sort out spam emails, even if the sender or the content is not yet known. Humans only intervene to check and correct, for example, if an email was incorrectly marked as spam.
Speech recognition systems are also capable of learning. The more you use them, the more they adapt to the user’s linguistic idiosyncrasies. In this way, they can better understand his voice over time and make fewer processing errors.
Automatically into space
In 1997 machines traveled to the planet Mars in the service of man. The aim of the “Pathfinder mission” was to bring scientific measuring devices to the surface of Mars. Appropriate techniques for the flight phase, entry into the atmosphere, descent, and landing should be developed and tested.
But the “Pathfinder mission” was successful and laid the foundation for further Mars missions. In August 2012, the vehicle “Curiosity” landed on Mars: 900 kilograms in weight and equipped with a variety of instruments to explore the