Why can AI pose a human threat?
byA Health Blog
Artificial intelligence (AI) has developed rapidly in recent years, and Google developed AI is GoDefeat the top players, Generate an episode of a fictitious characterPodcasts shedAnd, we have achieved amazing evolution. Some researchers and engineers around the world who advocate the danger of development of AI technology have also appeared, and about the question "Why is AI considered to be a human threat?", Scientific news Online site to reportLive ScienceIt is summarized.
Why Does Artificial Intelligence Scare Us So Much?
Electric car maker's "TeslaAnd "space project"SpaceXBusiness people who serve as CEO ofEarlon maskMr,"AI is more dangerous than North Korea"I've been ringing a warning on the dangers of AI for some time. Mr. Mask took place at the US Governor's Association meeting held in July 2017 towards attendeesAppeal that AI should be regulated, In AugustTo prevent arms race by autonomous homicide robotI am sending an open letter to the United Nations. "I have kept ringing alarm bells for a long time but people do not try to seriously think about AI until the AI loaded robot actually destroys the city and kills people," Mr. Mask , It says that we need to deal with AI before it brings harm to people.
Also, on March 14, 2018Passed awayTheoretical physicistStephen HawkingDr. at the conference held in 2015"Warning that AI exceeds humans within 100 years"And said that the evolution of artificial intelligence could bring about the end of mankind.
MIT scientistsNightmare machine"AI to create a nightmare-like image from a normal image" was created. The principal researcher who created the Nightmare machineManuel Cebrien"I am interested in how AI triggers emotions as fear," he said.
Cornell UniversityI will be Associate Professor of Computer Science DepartmentKillian Weinberger"A negative view towards AI is two patterns," AI itself tries to destroy humans with malice "and" patterns that immoral humans use AI for malicious purposes " I will think about it. "
"The idea that AI will treat humans as if humans now treat monkeys as a result of excessive development of AI is certainly undesirable," Wayneberger admitted. But Wayne Berger said that the idea that AI recognizes himself and causes treason to overthrow humans is based on a misunderstanding of AI. "AI will only work in a very limited range according to the algorithm that determines its behavior.AI can solve the problem faster than a human according to the algorithm for problems that AI is easy to deal with. However, AI is helpless against the many problems in the world, "Wayneberger said.
The development of AI technology certainly enabled AI to perform image recognition and processing at high speed and also realized to beat the top players of Go. However, regarding the point of "making AI conscious", it is said that research has not progressed to any extent. "I do not have any concern that AI will be conscious in the near future," Wainberger said.
Meanwhile, Wayneberger thinks that the possibility that "malicious people use AI" is much greater than the danger that AI is conscious.To divert AI to military technologyIs very effective, there is a risk that a malicious person gets a weapon equipped with AI and harm people. Mr. Wayneberger said, "In the military field, government regulation of AI is useful."
And if I could overcome the fear of AI,Diagnosis of disease by image recognitionWe can support AI to do it faster and more accurately, or look at the merits such as reducing accidents caused by human driving errors with automatic driving cars. It is likely that thinking how to effectively utilize AI not as a fearful thing as "a terrible thing" leads to future profits.