中国科技期刊使用人工智能政策的研究及对策建议OACHSSCD

Recently,artificial intelligence(AI)has developed rapidly.In the academic research field,AI has driven comprehensive innovation in research paradigms,theoretical modeling,and methodologies.In the academic publishing field,AI has reshaped the entire academic publication process,from manuscript writing,review,editing,and proofreading to dissemination.Moreover,AI has created a series of challenges for scientific research and academic publishing,such as the industrialization of"paper mills"and the fabrication of references.This urgently requires the academic publishing industry to reach a consensus;formulate clear norms and policies for the use of AI in academic publishing;and guide authors,reviewers,editors,and publishing units and other stakeholders to use AI reasonably and responsibly.This study selected 200 journals,including the top 100 natural science and social science journals in the 2024 Chinese Scientific and Technology Journal Citation Reports(Core Edition),as research objects and used the qualitative analysis tool Nvivo 12 Plus to analyze their AI-usage policy texts,aiming to provide references for Chinese scientific,technical,and medical(STM)journals in formulating AI usage policies and promoting their high-quality development.The number of reference points for author policies was the highest,totaling 531,followed by that for reviewer policies and editor policies,at 138 and 49 reference points,respectively.The number of reference points for publisher policies was the lowest,with only 29.Further analysis showed that regarding author policies,although Chinese STM journals have established tiered and categorized penalty mechanisms,there is a lack of consensus on requirements such as disclosure content,location,and statement templates.Regarding reviewer policies,Chinese STM journals clearly define the boundaries of using AI for review,but further clarification is needed on the responsibilities of reviewers and the disclosure requirements for AI usage.In terms of editor policies,Chinese STM journals have focused on regulating the scope of editors'prohibition of using AI but lack clear guidelines on the reasonable use of AI.In terms of publisher policies,Chinese STM journals lack comprehensive AI review and assessment mechanisms and a disclosure mechanism for AI usage.To this end,drawing on the practical experience of domestic and foreign STM journal publishing units and academic publishing organizations,four specific measures are proposed from the perspectives of authors,reviewers,editors,and publishing units as stakeholders.First,a disclosure mechanism for authors' use of AI is established from three dimensions:disclosure requirements,disclosure location,and content.Second,the principles of accountability and transparency should be adhered to,and the responsibility attribution and disclosure requirements for reviewers'use of AI should be clarified.Third,"human-machine collaborative"editing and publishing mechanism should be constructed by clarifying the boundaries of rights and responsibilities for AI use,establishing an infrastructure for AI-assisted editing and proofreading,and strengthening team development.Fourth,publishing units should establish AI usage policies covering the entire academic publishing process,forming a closed loop of"usage norms-review mechanisms-accountability mechanisms".