Who Shapes AI? Technology, Institutions, and Power in the Development of Artificial Intelligence

Abstract

Artificial intelligence is usually mainly explained through technological breakthroughs, such as larger models, improved algorithms, stronger computing power and expanded datasets. This article challenges this interpretation by viewing artificial intelligence as a sociotechnical system whose development and consequences are constantly influenced by social institutions. Based on research of the history of the Internet and human-computer interaction, institutional development, automated responsibility, and algorithmic labor, this paper compares early computing systems with contemporary artificial intelligence. Analysis shows that technical capabilities themselves cannot determine how a technology develops or how it affects society. Governments, universities, companies, markets, legal systems, workers and users all influence the way artificial intelligence is funded, deployed, regulated and understood. Meanwhile, these actors do not have equal power. Cases such as moral collapse zones and algorithmic wage discrimination demonstrate how automation unevenly redistributes responsibility, information and economic benefits. Therefore, this article holds that the future of artificial intelligence is not determined by the technology itself, but rather by the institutional structure and power relations in which the technology is developed and utilized.