Big Tech and Power in Artificial Intelligence Platforms: A Critical Review

Abstract

Critical, administrative and technical Communication and Social Sciences studies diagnose closed relationships between Big Tech and the new Artificial Intelligence summer. This consensus is synthetized with the idea of a Big AI and supposes that subsymbolic models function under the premise “bigger is better.” However, there are few analyses able to systematize the main arguments in this tacit agreement. Therefore, this paper aims to reconstruct the most significant arguments that highlight force relations in Big AI and group them into three areas of discussion. To analyse the corporate power within predictive and generative AI, this study is grounded in a methodology composed by academic review and documentation analysis. This approach, firstly, will allow recovering studies on market capitalization dynamics, political-institutional dimensions, and geopolitical conditions that coagulate the Big Technification of AI. Then, the study will be able to explore researches on the industrialization processes of AI capitalism and its value chains and app ecosystems. Finally, the paper will reconstruct studies on the ways in which platformization of machine and deep learning services mobilizes industrial power. With this reconstruction, the paper provides a synthesis of usual arguments that evaluate power within foundation models multiplication, but also delimitates frontiers for conceptualising power relations in AI pervasiveness. Furthermore, these themes will provide a framework for discussing gaps linked to the need for situated studies that explore the entire technology stack and examine the infrastructural power of AI.