White Robin R
School of Animal Sciences, Virginia Tech, Blacksburg, VA 24060.
JDS Commun. 2025 Jan 6;6(3):452-457. doi: 10.3168/jdsc.2024-0707. eCollection 2025 May.
The launch of generative artificial intelligence (GenAI) tools has catalyzed considerable discussion about the potential impacts of these systems within the scientific article preparation process. This symposium paper seeks to summarize current recommendations on the use of GenAI tools in scientific article preparation, and to provide speculations about the future challenges and opportunities of GenAI use in scientific publishing. Due to the dynamic nature of these tools and the rapid advancement of their sophistication, the most important recommendation is that ongoing engagement and discussion within the scientific community about these issues is critical. When using GenAI tools in scientific article preparation, humans are ultimately accountable and responsible for products produced. Given that accountability, an expert panel convened by the National Academies of Science, Engineering, and Medicine recently proposed principles of GenAI use in science communication, including (1) transparent disclosure and attribution; (2) verification of AI-generated content and analyses; (3) documentation of artificial intelligence (AI)-generated data; (4) a focus on ethics and equity; and (5) continuous monitoring, oversight, and public engagement. In addition to the importance of human accountability, many publishers have established consistent policies suggesting that GenAI tools should not be used for peer reviewing, figure generation or manipulation, or assigned authorship on scientific articles. Along with the potential ethical challenges associated with GenAI use in scientific publishing, there are numerous potential benefits. Herein we summarize example conversations demonstrating the capacity of GenAI tools to support the article preparation process, and an example standard operating procedure for human-AI interaction in article preparation. Finally, diverse broader questions about the impact of GenAI tools on communication, knowledge, and advancement of science are raised for rumination.
生成式人工智能(GenAI)工具的推出引发了关于这些系统在科学论文撰写过程中潜在影响的大量讨论。本研讨会论文旨在总结当前关于在科学论文撰写中使用GenAI工具的建议,并对GenAI在科学出版中的未来挑战和机遇进行推测。由于这些工具的动态性质及其复杂性的快速提升,最重要的建议是科学界就这些问题持续参与和讨论至关重要。在科学论文撰写中使用GenAI工具时,人类最终要对所产生的成果负责。鉴于这种责任,美国国家科学院、工程院和医学院召集的一个专家小组最近提出了在科学传播中使用GenAI的原则,包括:(1)透明披露和归属;(2)对人工智能生成的内容和分析进行验证;(3)记录人工智能生成的数据;(4)关注伦理和公平;(5)持续监测、监督和公众参与。除了人类责任的重要性外,许多出版商已经制定了一致的政策,表明GenAI工具不应被用于同行评审、图表生成或处理,或在科学论文上指定作者身份。除了与在科学出版中使用GenAI相关的潜在伦理挑战外,还有许多潜在的好处。在此,我们总结了一些示例对话,展示了GenAI工具支持论文撰写过程的能力,以及一个在论文撰写中人类与人工智能交互的示例标准操作程序。最后,还提出了关于GenAI工具对科学传播、知识和进步的影响的各种更广泛的问题以供思考。