Expert opinion: AI is still very "weak", how can it compete with humans?

[51CTO.com original article] "I am neither a machine nor a human being, I am better than both." This is a classic line from the 2015 movie "Terminator: Genisys" starring the famous American actor Arnold Schwarzenegger. During the boom of artificial intelligence technology, some people have also raised their concerns - are they safe? Will humans be replaced by them? Especially after the man-machine battle between Lee Sedol and AlphaGo this year, when the total score was fixed at 1:4, such doubts became even stronger.

[[179054]]

Is this kind of argument or panic necessary? Let's answer this question from the side:

1. In February 2015, a 14-year-old boy in the United States successfully hacked into a connected car, and the tools he used were only $15 worth of electronic parts. The boy bought the required electronic parts at the electronics retailer RadioShack for $15 and made a circuit board in one night. The next day, he successfully hacked a car's security system with an iPhone and this homemade device.

2. In July 2015, according to the Associated Press, Volkswagen said on Wednesday that a skilled worker was killed by a robot he installed in one of its German factories. Volkswagen spokesman Hilwig said that the preliminary conclusion was that it was caused by human error and there was no problem with the robot.

3. At the GeekPwn2016 Carnival on October 24 this year, Stephen Chavez, who is known as the "Little Hawking" in the field of Internet security and suffers from schizencephaly but has 10 years of programming experience, was the first to appear. He demonstrated how to use vulnerabilities to remotely control a stranger's smart wheelchair. At the same time, the Changting Technology Team from Tsinghua University also demonstrated jailbreaking for the PS4 game console.

It can be seen from a few simple incidents that whether it is robots "killing people" or smart terminals being controlled, they are all caused by human factors. With the current level of technology, it is completely impossible for artificial intelligence to have autonomous consciousness without human control.

However, concerns about artificial intelligence are normal. But history has proven that most of the past economic booms were caused by the adoption of new technologies. Of course, artificial intelligence still has a long way to go. Judging from the current speed of development, although recognition programs and voice recognition have made great progress, robots perform poorly in open and chaotic environments. The biggest obstacle that artificial intelligence ultimately faces is that "machines are machines after all." From the current state of technological development, it is impossible for artificial intelligence to defeat humans. If it really happens, it will be the humans who control artificial intelligence who defeat humans. Even the movie "Terminator" reflects this point. "Skynet" was designed and launched by humans. "Judgment Day" is not so much the rise of robots as it is the self-destruction caused by humans' overconfidence in the technology they have developed.

Academician Li Deyi also mentioned in his speech titled "Artificial Intelligence is Running" at the Global Artificial Intelligence Technology Conference on April 22 that "the AlphaGo program still has many shortcomings. When it plays chess, it needs an assistant to hold the chess pieces. Its hands are not yet made, it has no eyes, and no sense or behavioral ability. This robot must have its perception and behavioral abilities, and it is not enough for it to have only thinking ability. In addition, the AlphaGo program currently has no emotions and feelings, and cannot analyze the opponent's psychological state on the spot. It is the same no matter who it plays with. It cannot engage in psychological warfare with the opponent on the spot and lacks the ability of interactive cognition."

As for those arguments about the job crisis brought about by the development of artificial intelligence, there is even less need for them. Objectively speaking, all simple and repetitive mental labor will be replaced by artificial intelligence in the future. However, the development of artificial intelligence may bring some surprises, just like the first steam car made by the Frenchman Giunault in 1769. Although the first test run caused the first motor vehicle accident in human history, it did not stop the research and development of automobiles. As for the worry that artificial intelligence will replace humans, it is actually unnecessary - when old occupations and positions disappear, new positions and even new occupational forms will emerge. But it is undeniable that when the era of artificial intelligence really arrives, there will be only two things in the world, humans and intelligent robots. Because all devices are "humanized". Of course, intelligent people look very different, either computers or cars.

Although artificial intelligence has been developed for 60 years, it is still a "new thing" for us. Unknown or poorly understood things always bring curiosity and fear to humans. What impact may artificial intelligence have on humans? It is difficult to determine at present, but the threat theory has always accompanied its development. It is obviously not enough to just study the technology of artificial intelligence. The technology has developed ahead of time. American scholars believe that artificial intelligence may be the only technological development that humans cannot try and fail multiple times or even just once.

As for the future development of artificial intelligence, based on the current development situation we know, it will bring some "harm" to humans in the future, such as unemployment in a certain period, human injuries due to technical and operational problems, etc. However, these are not reasons to hinder the development of artificial intelligence. History has proved that humans will not stop moving forward in the face of new things because of fear. Therefore, we should face artificial intelligence with a more positive attitude. With the improvement of big data, machine learning and computing power, the development potential of artificial intelligence is unlimited, and it will eventually bring more services and convenience to human society.

[51CTO original article, please indicate the original author and source as 51CTO.com when reprinting on partner sites]