Source: China Publishing Journal

Driven by the wave of digitalization, artificial intelligence (AI) is penetrating the entire academic publishing chain at an unprecedented pace, bringing about a major technological overhaul. It has not only lifted efficiency across the board but also redefined business models in the publishing value chain. However, rising risks such as algorithmic bias, academic fraud, and data breaches have put academic editors' gatekeeping function under new pressure.

Improving content production efficiency. It is now an industry consensus that AI can significantly boost efficiency across the entire content production chain of academic publishing. In topic planning, leveraging big data mining and topic prediction models, AI can automatically identify academic hotspots, rapidly retrieve vast volumes of literature for data analysis and model development, significantly shorten research cycles, and support topic selection decisions by accurately pinpointing high-value research directions. In content presentation, AI-generated content (AIGC), with its robust generative capabilities, can automatically draft manuscripts and diversify content presentation formats. In knowledge acquisition, AI-powered retrieval and knowledge association technologies, combined with question-driven reading, enable readers to engage more actively in knowledge acquisition and enhance their reading experience. In content review and proofreading, intelligent peer review systems integrated with comprehensive academic literature databases can assist in plagiarism detection and conduct preliminary screening of submissions, thereby providing references for editorial decisions. In reviewer matching, data analysis of paper characteristics and reviewer profile databases enables more efficient matching of appropriate reviewers. In typesetting support, AI-driven automatic typesetting and page layout optimization significantly reduce formatting-related costs.

Addressing AI-related academic risks. While academic publishers embrace the convenience brought by new technologies, it is imperative to clearly define the boundaries of technological application. At present, the threats posed by AI to academic publishing are mainly reflected in the following four aspects: 

First, algorithmic bias and the resulting false information may trigger an academic integrity crisis. The locality of training data poses severe challenges to academic objectivity and fairness. In content production, AI may generate misleading content and so-called “AI hallucinations”, thereby eroding public trust in academic research.

Second, imbalances between data privacy protection and intellectual property rights protection are becoming increasingly prominent. As AI becomes more deeply integrated into content production, the use of unauthorized training corpora by large models may lead to a series of issues, including blurred legal boundaries, exorbitant rights protection costs and minimal piracy costs. This has become a core risk point in the publishing value chain.

Third, excessive reliance on automation will weaken researchers' innovative capacity and critical thinking. The deep intelligentization of the entire academic publishing process will inevitably lead to over-dependence on technology. Since AIGC is trained on existing databases and lacks the uniquely human capacity for intellectual insight and epiphany, over-reliance on AI will stifle innovation and value creation in academic research, erode editors’ decision-making authority, and marginalize professional judgment in academic gatekeeping.

Fourth, the high barriers to AI development risk creating monopolies in the academic publishing industry. Currently, the publishing industry is generally confronted with the “three highs” challenge in large model application: high access costs, high computing power requirements and high development complexity. In the future, the widening digital divide could make the emergence of publishing monopolies inevitable.

The restructuring of publishing production order triggered by technological transformation is also a process through which publishers reassess the value of academic publishing. Rational and prudent understanding and application of new technologies, and effectively promoting the high-quality production and dissemination of academic achievements, constitute core issues concerning the survival and development of academic publishing. Only under an integrated framework encompassing technology, governance, business, and talent can academic publishing achieve sustainable development characterized by improved efficiency, enhanced value, and controllable risks. Against this backdrop, academic editors should take the initiative to grasp the underlying principles and future trends of academic publishing, uphold their central role throughout the publishing process amid both change and continuity, and consistently adhere to the ultimate goals of pursuing truth and knowledge innovation, maintaining rigorous and evidence-based research traditions, and emphasizing problem-oriented inquiry and practical investigation. On this basis, editors should continuously develop new skills and transform from traditional “craftsmen of text” into “knowledge service providers” in the digital and intelligent era.