Critical Theory of AI: The Alienation of the Human Intellect

In this article, I present some fundamentals of a critique of AI influenced by the Critical Theory of the Frankfurt School. From this perspective, artificial intelligence represents a kind of apotheosis of instrumental rationality, an alienation of the human intellect, framed by economic calculations, geopolitical strategies, the reduction of reason to cognition, perception to classification and knowledge to an inventory. The article begins by defining machine-learning-based forms of artificial intelligence, distinguishing between different types of foundation model (LLM-based and non-LLM-based), describing how LLM-based GenAI works and discussing the realistic prospects for the emergence of AGI in the near future, especially when this is anticipated from increases in compute scale of GenAI. Then, it outlines the fundamental critical position of Frankfurt School Critical Theory, which is that human knowledge cannot be reduced to an inventory of items any more than the human intellect can be reduced to instrumental rationality. Against this background, it critiques the application of AI as a business machine and a warfare technology, arguing that it is crucial, in opposing corporate and military applications, not to uncritically accept the reduction of the human intellect to mental labor. Nonetheless, the implications of AI for the automation of mental labor are critically important, and so the article turns next to the microeconomics of AI implementation, inferring from the discourse of liberal economics that the disruptive structural transformation of the middle class through polarization and casualization is a real possibility. In conclusion, the article discusses the implications of this dialectical critique, not just in terms of the potential for a political convulsion, but also in terms of the urgent need for AI research to acknowledge (and address) the problems and limitations of current foundation models.