In the face of the unprecedented "artificial intelligence tide" in the scientific and technological circles, technology giants such as Musk and Gates can't sit still. 20/kloc-in August, 2004, Musk posted a status on Twitter: "The book Super Intelligence written by bostrom is worth reading. We must be super cautious about artificial intelligence, and its potential may be more dangerous than nuclear bombs. " Bill? Gates also said in a recent activity that machines will replace human beings in various jobs, and if we can handle them well, they should be able to play a positive role. However, in a few decades, artificial intelligence will develop to a worrying level.
The hidden danger of artificial intelligence: or the impact on the existing social system
At the beginning of February, hundreds of top scientists at the Future of Life Institute, which is dedicated to "reducing the existing risks faced by human beings", signed an open letter, calling on the scientific and technological community to thoroughly examine the impact of artificial intelligence development on human society while promoting the development of artificial intelligence.
The author of this letter admits that artificial intelligence has achieved remarkable success in the fields of speech recognition, image classification, self-driving cars, machine translation and question answering system, and believes that these studies are helpful to eliminate disease and poverty. But they insist that "artificial intelligence system must do what we want it to do" and the scientific community should establish the research goal of "helping to maximize the social benefits of artificial intelligence".
The open letter emphasizes that artificial intelligence may affect society in the fields of computer security, economy, law and philosophy in the future, but it is this ubiquitous potential that can trigger various crises. For example, scientists believe that if some parts of the economy become automated in the future, workers and their wages will be greatly affected, and artificial intelligence experts have to pay attention to this. Artificial intelligence should face up to the following questions: Does lethal automatic weapons conform to humanitarian law? When the artificial intelligence system collects data from surveillance cameras, telephone lines and emails, how to avoid invasion of privacy?
A group of scholars from Oxford University wrote in their blog last year: "When a machine makes a mistake, its error mode may be more dramatic than human error, which may bring more unpredictable results. A simple algorithm is largely predictable, but it may make strange decisions under unusual circumstances. "
Compared with these practical problems, Hawking, a well-known scientist, takes a longer view. Hawking said in an interview with the BBC that if technology has similar capabilities to humans, "it will get out of control and redesign itself with acceleration. What is even more frightening is that due to the limitation of biological significance, human beings can't keep up with the development speed of technology. " Hawking expressed his concern about the future of mankind: "Due to the limitation of slow biological evolution, human beings cannot compete with machines and will be replaced. The development of all artificial intelligence may lead to the end of mankind. "
From the comments of these scientists, it can be seen that the concerns of scientific and technological circles about artificial intelligence mainly come from two aspects: on the one hand, the changes of artificial intelligence on human work and lifestyle may affect the existing legal system, moral standards and interest distribution methods. And the speed of human change may not keep up with the development speed of artificial intelligence, which will impact the existing system of society and lead to chaos; On the other hand, artificial intelligence makes people think less and less, and at the same time it may become smarter and smarter, thus threatening human survival.
Predictions come true? The "destruction theory" is somewhat alarmist.
For the remarks of Musk, Gates, Hawking and others, Google Chairman Schmidt took the lead in launching a rebuttal. Schmidt has a special identity. As the chairman of Google, he personally participated in the research and development of many of the most complex artificial intelligence systems in the world, from self-driving cars to Google's predictive search engine. Google even launched its own in-house robotics lab last year. Schmidt believes that all the fears that machines will take away human work and occupy the world are groundless. I hope people can realize that robots will become our friends.
Schmidt said: "It is normal to worry about artificial intelligence. But looking back at history, you will find that these are absolutely wrong. We just hope to put on better clothes through more mechanized methods. Historical experience has proved that most of the past economic prosperity was triggered by the adoption of new technologies. "
Wang Yang, vice president of IBM, once told Tencent that when technology has made rapid progress, people will always feel scared, but what really matters is the people who master the technology. "There are always villains making trouble with new technologies", and the development of technology is unstoppable. For example, in Hollywood blockbusters, evil forces often abuse high-tech technology, but human beings must master these as soon as possible to resist the invasion of unjust people. In a word, the development of new technology is bound to be controversial, but the progress of science will always bring a better future to mankind. "
Tony? Cohen is a professor of automatic reasoning at Leeds University. He said that complete artificial intelligence "still has a long way to go. Judging from the current development speed, I think it will take nearly a hundred years. " Cohen believes that although great progress has been made in recognition programs and speech recognition, robots do not perform well in an open and chaotic environment, and the biggest obstacle facing artificial intelligence in the end is that "machines are machines after all."
In the eyes of most people, the assertion that "artificial intelligence destroys mankind" seems to be somewhat exaggerated. The existing technology can only make the machine engage in some basic things, such as some fixed actions on the production line, and even the intelligent system on the driverless car is not perfect, not to mention the artificial intelligence with thinking and emotion, which can pose a threat to human survival. However, one thing cannot be ignored. With the progress of science and technology, automated artificial intelligence will appear more and more in our work and life, and it is very likely that in the near future, machines will replace human work.
Driverless car is the nearest artificial intelligence to us, but it faces many challenges besides technical factors. For example, Google's driverless car had a car accident, which caused serious consequences. Is the owner, Google, or the algorithm, sensor and all control systems driving the car responsible for the traffic accident? Unfortunately, the existing legal system has failed to keep pace with the development of modern society, enterprises or robots, and it is impossible to give a clear answer.
We should also think about what we should do when robots make a large number of workers unemployed. When the robot misoperation causes an accident, how does the existing law clarify the responsibility? When hackers use artificial intelligence to commit crimes, how should we guard against it? These should be the problems we need to face up to now.