The Soul and the Algorithm: Rethinking Peer Review in the Age of Artificial Intelligence

Peer review constitutes the bedrock of quality assurance and integrity in scientific research. However, the traditional system, inherently reliant on human reviewers, faces growing challenges such as workload burden, inconsistencies, and biases. Concurrently, the advancement of Artificial Intelligences (AIs), notably Large Language Models (LLMs), raises discussions about their potential to optimise or transform this process. This article analyses the limitations of human-led peer review and explores the contributions and current deficiencies of AIs in this domain, including the origin of biases in their training data and the need for criticism prioritisation. It is argued that, while AI offers promising capabilities in efficiency and access to information, it still lacks the nuances, critical judgment, and prioritisation ability intrinsic to human competence. Through the presentation of solutions for mitigating human and AI limitations, and a simulated case study, this work suggests a future where synergistic collaboration between AI and human reviewers can optimise the quality of academic assessment, transforming current challenges into opportunities for continuous improvement. Keywords: Peer Review; Artificial Intelligence; Human Limitations; AI Limitations; Academic Publishing.