Nick Bostrom, director of the Institute for the Future of Humanity at Oxford University, lays out the research community's best predictions about artificial intelligence (AI) in his new book, Superintelligence: Paths, Dangers, Strategies.
Here are the combined results from four surveys by AI researchers, including a survey of the most cited scientists in the field, totaling 170 respondents.
10% believe that the intelligence of robots will be equal to the human mind by 2022; 50% believe that this will happen in 2040, and 90% agree that robots will become smarter than humans by around 2075.
Machine intelligence at the human level is defined here as "one that can perform most human occupations - at least those typical of humans."
By this definition, we shouldn't be surprised at such predictions. Robots and algorithms are already squeezing the global workforce market. Routine jobs are becoming digitalized: farmers, telemarketers, stock traders, credit staff, lawyers, and journalists are all already feeling kicked in the back by their new automated counterparts.
Bostrom does not consider the replication of routine to be intellectual activity. He talks about intelligence in combination with intuition and logic - this type of mind can cognize and feel the world around, coping with its uncertainty. The most interesting thing in achieving the human level of intelligence is not the achievement itself, but what will happen next. Once machines can reason and improve themselves, only SkyNet can be their limit.
Promotional video:
In this logic, superintelligence is defined as one that "significantly exceeds human cognitive performance in almost all areas of interest."
10% of scientists believe that superintelligence will appear in a couple of years; 75% of respondents believe that this will happen in 30 years.
Tesla inventor and CEO Elon Musk warns that superintelligent machines are arguably the greatest threat to human existence. “I really hope we are more than just a biological downloader for digital superintelligence, which is unfortunately becoming more and more likely,” Musk says.
Nevertheless, the best minds in these technologies predict the end of human intellectual superiority by the middle of this century.
"The success of AI will be the biggest event in human history," writes theoretical physicist Stephen Hawking, "but it could also be the last event in human history."