What is a friendly AI?
friendly AI is the term used by scientists of artificial intelligence (AI) to refer to AI systems that generally perform actions that are useful to humanity than neutral or harmful. This does not mean blind obedience - on the contrary, this term is used to describe AI systems that are friendly because they want to be, not because of any externally stored force. In addition to referring to completed systems, friendly AI is also the name of the theoretical and engineering discipline, which would be used to successfully create such systems.
The term "friendly AI" was created with Eliezer Yudkowsky Singularity Instate for artificial intelligence, which aims to create a friendly AI smart enough intelligent AI. His work with a long book on this topic, creation of friendly ai , published online in 2001, is probably the first strict treatment of the topic anywhere. Yudkowsky evokes the argument of evolutionary psychology and other cognitive sciences after supporting his approach to probesa friendly AI.
friendly AI is considered important as academic discipline, because the past is trying to "answer" the problem of dishonest AI generally induced strict program restrictions that must be collapsed in alternative interpretations when AI becomes smarter than humans, or simply when gaining the ability to reprogram. Antropomorphism is also a problem of AI. Because evolution builds organisms that tend to be selfish, many thinkers assume that any AI we build would have the same tendency, either immediately or after smart enough.
evolution creates organisms with focused target systems because there is no other way. Altruistic target systems took many millions of years to evolve and did so only under conditions in which members of the same HA Had tribe a lot by helping each other and losing a lot ofEdbing. But if we were to design the mind from scratch, we could build it without the target target system. This would not be "ai restriction" - it would simply create AI, which is unobtrusive from the nature.
The above observation is one of many that contribute to the area of friendly AI, which is extremely new and still needs a lot of work. Some scientists AI claim that we cannot determine the design functions necessary to implement a friendly AI until we have smarter AI systems on which we will experiment. Others argue that a purely theoretical approach is not only possible, but ethically necessary before the start of a serious attempt on the General AI.