In Future Progress in Artificial Intelligence: A Survey of Expert Opinion, by Vincent C. Müller and Nick Bostron, the authors surveyed 170 Artificial Intelligence experts to see if they believe whether or not there will be an “intelligence explosion” in the future. 

The authors defined “intelligence explosion” as “a growth beyond human ability and perhaps even an accelerating rate of growth.” The pair wanted to know the impact that this superintelligence would have on society and when this explosion would be occurring. 

I found this research article to be quite confusing. At the end the authors informed the readers that the “results should be taken with a grain of salt” and that the reader would be responsible for drawing their own conclusions from the survey data. I understand that there is a need for additional studies to be done but I saw many flaws in this research. For example, the authors said they often had to “pressure them via personal email to respond.” I question whether “pressuring” participants is really an effective way to get them to respond to the questionnaire. Many of the respondents also called out the authors for making a “biased” questionnaire. The only conclusion that the pair drew from the questionnaire was that they expect that AI will most likely reach 50% of human abilities by 2040-2050 and 90% by 2075. Lastly, they saw no circumstances in which superintelligence could be good for humanity.