This article presents Nick Bostrom’s considerations of the future included in his book Superintelligence: Paths, Dangers, Strategies. Bostrom studies such issues as determining the hypothetic ways of attaining superintelligence, its nature and different aspects of this technology. He shows threats regarding such powerful systems, as well as constructing strategies of preventing undesirable activities of superintelligent beings. Bostrom’s input is an important part of present discussion concerning the development of artificial intelligence and its ethical problems.