The creation of facts in the cloud

Abstract

Like most significant changes in technology, Cloud Computing and Big Data along with their associated analytic techniques are claimed to provide us with new insights unattainable by any previous knowledge techniques. It is believed that the quantity of virtual data now available requires new knowledge production strategies. Although they have yielded significant results, there are problems with advocated processes and resulting facts. The primary process treats "pattern recognition" as a final result rather than using "pattern recognition" to lead to yet to be tested testable hypotheses. In data analytics, the discovery of a pattern is treated as knowledge rather than going further to understand the possible causes of those patterns. When this is used as the primary approach to knowledge acquisition unjustified inferences are made - "fact generation". These pseudo-facts are used to generate new pseudo-facts as those initial inferences are fed back into analytic engines as established facts. The approach of generating "facts from data analytics" is introducing highly risky scenarios where "fiction becomes fact" very quickly. These "facts" are then given elevated epistemic status and get used in decision making. This, misleading approach is inconsistent with the moral duty of computing professionals embodied in their Codes of Ethics. There are some ways to mitigate the problems generated by this single path approach to knowledge generation.