Google's Expenditures recognized as shon as scary as nuclear weapons: what is it dangerous

Didzhital" on Facebook according to Eric Schmidt, in the near future because of problems with China and the US security will have to conclude an agreement on restraint in the field of artificial intelligence, just as previously forbade the test of nuclear bombs and signed others Important contracts in this area. In the 1960s, the countries created a world that made it impossible to "surprise" and reached the balance of trust.

Now no one is working on such restrictions for artificial intelligence, although it is "just as powerful". "I am very concerned that the US attitude to China as a corrupt, communist or another country, and Chinese attitude to America as a failure will allow people to say," My God, they think something "and then launch something like a puzzle . Starting to do something for weapons or preparation, you cause concern on the other party, " - said Eric Schmidt.

According to Schmidt, he was not fully aware of the power of information when Google headed. Now he sees the potential harm of technology, so he calls for them to create them taking into account the ethical norms and moral values ​​of humanity. Hiroshima and Nagasaki are two Japanese cities that the United States destroyed at the end of World War II by dropping nuclear bombs. This attack claimed the lives of tens of thousands of people and showed how powerful and scary nuclear weapons are.

After that, the USSR and China also mastered technology, and humanity lives under threat of guaranteed mutual destruction. Restrain theory implies that if one country launches nuclear warheads, then they will be answered the same, and a global disaster will occur. Vice notes that AI and machine learning are striking technologies that often understand the wrong.

For the most part, machines are not as smart as people think, but quite capable of creating masterpieces of art, winning video games or communicating with clients on the phone. Some developers are trying to force artificial intelligence to drive a car in large cities without human participation, but there are still cases. Others suggested that you have control of nuclear weapons, believing that it would not make mistakes and help to avoid catastrophe.

"AI is far from as destructive as nuclear weapons. But people who are in power are afraid of this new technology and, as a rule, for completely wrong reasons," analysts say. They explain that the main problem of artificial intelligence is not the potential that can be compared to nuclear weapons, but the disadvantages of developers. Machine algorithms are good as much as the people they create.

For example, racists, consciously or not, convey their views on chatbots, which then allow racist statements. DEEPMIND DEPMIND CEO Hasabis, who created AI, capable of winning people in Starcraft II computer game: "AI, is a very large -scale thing, who creates the si, which cultures they come from and what values ​​in them, the creators of the systems of the AI.