A Criticism of the Technological Singularity


The compelling story of self-improving machines which become exponentially smarter up to inconceivable intelligence has inspired some of the best science fiction literature [1–3], but is also taken seriously by many researchers. This story is however based on empirical observations of seemingly exponential processes such as Moore’s law in the semiconductor industry, and contains multiple fallacies concerning self-improvement of intelligent systems (including humans), which upon close look are implausible. Deep Learning has been heralded as a major step in this direction, however a closer look again shows many open issues with this approach, leading us to conclude that we deciphered only a small part of one method which nervous systems may use to create intelligent behaviour; that seemingly simple tasks like image classification and segmentation are still AI-complete; and that true Artificial General Intelligence (AGI) still lies at least several centuries in the future. But even an AGI would not be able to exponentially self-improve without further advances. These fallacies abound in science fiction literature as well as in scientific papers, and we will illustrate our analysis with appropriate examples.

Seewald A.K. (2022) A Criticism of the Technological Singularity. In: Dingli A., Pfeiffer A., Serada A., Bugeja M., Bezzina S. (eds) Disruptive Technologies in Media, Arts and Design. ICISN 2021. Lecture Notes in Networks and Systems, vol 382. Springer, Cham.