What is the Technological Singularity?

The Technological Singularity, or simply "Singularity," is a multi-faceted concept in futurism with several overlapping and sometimes conflicting definitions. The most proper and prominent definition of the Singularity was given by Vernor Vinge in his essay, The Coming Technological Singularity. It refers to the point at which superhuman intelligence is created technologically. These superhuman intelligences could then apply their brainpower and expertise to the task of creating additional or more powerful superhuman intelligences, resulting in a snowball effect with consequences beyond our present ability to imagine.

The term "Technological Singularity" was coined by analogy to the singularity in the center of a black hole, where the forces of nature become so intense and unpredictable that our ability to calculate the behavior of matter in these circumstances drops to zero. Often mentioned in conjunction with the idea of superhuman intelligence in Singularity dialogues is the notion of accelerating technological change. Some have argued that as the slope of technological progress increases, it will culminate in an asymptote, similar visually to a mathematical singularity.

However, this notion of the singularity is not the same as Vinge intended; referring to the emergence of superhuman intelligence, along with superhuman thinking speeds. (Including smartness, ability to understand and create concepts, turn data into theories, make analogies, be creative, and so on.) Though superhuman intelligences creating additional superhuman intelligences would indeed result in the acceleration of technological progress, progress would not become infinite, in the sense that a mathematical singularity would suggest.

Because superhuman intelligences would, by definition, be smarter than any human, our ability to predict what they would be capable of with a given amount of time, matter, or energy are unlikely to be accurate. A superhuman intelligence might be able to fashion a functioning supercomputer out of cheap and readily available components, or develop full-fledged nanotechnology with nothing but an atomic force microscope. Because the ability of a superhuman intelligence to design and manufacture technological gadgets would rapidly surpass the best efforts of human engineers, a superhuman intelligence could very well be the last invention that humanity ever needs to make. Due to their superhuman genius and the technologies they could rapidly develop, the actions of intelligences emerging from a Technological Singularity could result in either the extinction or the liberation of our entire species, depending on the attitudes of the most powerful superhuman intelligences towards human beings.

Oxford philosopher Nick Bostrom, director of the Oxford Future of Humanity Institute and the World Transhumanist Organization, argues that the way superhuman intelligences treat humans will depend on their initial motivations at the moment of their creation. A kind superhuman intelligence would, in wanting to preserve its kindness, beget kind (or kinder) versions of itself as the self-improvement spiral continued. The result could be a paradise in which superhuman intelligences solve the world's problems and offer consensual intelligence enhancement to human beings. On the other hand, a malicious or indifferent superhuman intelligence would be likely to produce more of the same, resulting in our accidental or deliberate destruction. For these reasons, the Technological Singularity might be the single most important milestone our species will ever confront.

Several paths to superhuman intelligence have been proposed by Singularity analysts and advocates. The first is IA, or Intelligence Amplification, taking an existing human and transforming her into a nonhuman being through neurosurgery, brain-computer interfacing, or perhaps even brain-brain interfacing. The other is AI, or Artificial Intelligence, the creation of a dynamic cognitive system surpassing humans in its ability to form theories and manipulate reality. When either of these technologies will reach the threshold level of sophistication necessary to produce superhuman intelligence is uncertain, but a variety of experts, including Bostrom, cite dates within the 2010-2030 range as likely.

Because the Singularity may be nearer than many would assume, and because the initial motivations of the first superhuman intelligence may determine the fate of our human species, some philosopher-activists ("Singularitarians") view the Singularity not only as a topic for speculation and discussion, but as a practical engineering goal that meaningful progress can be made towards in the present day. Thus, in 2000 the Singularity Institute for Artificial Intelligence was founded by Eliezer Yudkowsky to work exclusively towards this goal.