Orchestrating Critique in the Era of Large Language Models (On Scientific Review and Peer Review in the Age of Artificial Intelligence: A Framework Proposal)

This paper examines the fundamental transformations in the concepts of critique and scientific peer review in the era of large language models. By introducing the concept of transition from "print human" to "post-print human," the author demonstrates how the emergence of artificial intelligence has blurred traditional boundaries between human and machine agency in the production and evaluation of meaning. The paper first explores the notion of "independent authorship" and the historical role of tools in the meaning-creation process, then shows how in the digital age, with the emergence of "cognitive assemblage" between human and machine, the traditional concept of scientific peer review faces a crisis. Using the metaphor of river and mill, the author explicates the inherent limitations of language models—including their inability to grasp "significance," structural biases, and generation of "plausible falsehoods." Finally, a four-stage framework for "orchestrating critique" is proposed, comprising: machine agent recognition, human agent skepticism, humanizing cognition, and meta-evaluation of evaluation itself. This framework is built on "constructive distrust" and requires establishing an institutional system with evolutionary collective memory capabilities and preservation of epistemological diversity, transforming the reviewer from a solitary critic into an orchestrator of a complex network of meaning production.