Where is artificial intelligence on the road "super smart"?

Robotics> Robot Web: If you understand various viewpoints about artificial intelligence, you may be very worried about "super intelligence" or "killer artificial intelligence".

The MIT Technology Review recently published an article by Oren Etzioni, CEO of the Allen Institute of Artificial Intelligence and professor of computer science at the University of Washington, analyzing the possibility of super artificial intelligence. Accioni's research concluded that super intelligence will not appear in the foreseeable future.

The articles and reports published in the MIT Science and Technology Review are basically similar. The following is the full text of the article:

If you understand the various viewpoints about artificial intelligence, you may be very worried about "super intelligence" or the "killer artificial intelligence". The British "The Guardian" warned recently that "we are like children playing a bomb", and "Newsweek" pointed out that "artificial intelligence is coming and it may completely eliminate us."

Supported by the opinions of Elon Musk and Stephen Hawking, many media have reported this. In fact, this view comes from one person's research: British professor Nick Bostrom, author of the philosophical book Super Intelligence: Roads, Dangers and Strategies.

Bostrom is a philosopher at Oxford University, and the quantitative assessment of risk is the field of actuarial science. He may be considered the world’s first “actuary philosopher.” Considering that philosophy is a concept argumentation subject, and risk assessment is data-driven statistical work, such titles appear contradictory.

So, what does the data tell us? In 2011, as a member of the Greek Artificial Intelligence Association, Bosterrom attended the "Artificial Intelligence Philosophy and Theory" conference held in Thessaloniki, Greece. At the time, he collated the results of four investigations and gave the possibility of human brain level artificial intelligence appearing at different time points:

2022: 10%.

2040: 50%.

2075: 90%.

With regard to the development of human brain-level artificial intelligence, Bosterrom wrote a 300-page philosophical argument, and these four surveys are the main source of data.

In order to more accurately assess the opinions of leading industry researchers, I sought the help of the AAAI Fellows. These researchers have made important and sustained contributions to the research of artificial intelligence.

At the beginning of March 2016, AAAI sent an anonymous questionnaire to me and asked the 193 fellows the following questions: “In his book, Nick Bostrom asserts super intelligence as 'in various practical fields. , including scientific creativity, general common sense, and social skills, intelligence that is even smarter than the best human brain. "Do you think we will see super intelligence?"