Control of AI or Control over AI

Abstract

The technological development of mankind is significantly ahead of its institutional development. This necessitates the revision of existing means of institutional control over AI. Revolutions in the development of computer science (the flourishing of machine learning, artificial intelligence) and in the development of biology, in particular neuroscience open up new opportunities in the AI implementation and embodiment. Scientists invent biometric sensor as a direct link between AI and the human brain. This and other scientific discoveries make it necessary to consider the following questions: the risk of the emergence of digital dictatorships; the potential for AI to control or serve surveillance capitalism; AI as a possible defense against the misuse of AI. The methodology of this research is value analysis and institutional analysis. The aim of the article is to outline the current situation with the definition of possible value criteria for determining the goals of AI control and possible institutional ways of implementing such control. It is provided an assessment of the current effectiveness of existing institutional controls over AI.