Over 100 researchers gathered in Beijing on Sunday for the second Yuyuantan Conference on Research Integrity, discussing the responsible use of AI in academic publishing. The event released the Guideline on the Boundaries of AI-Generated Content Usage in Academic Publishing 3.0. Experts stressed verifying AI outputs to prevent errors and promoting global standards.
Over 100 researchers convened in Beijing on Sunday for the second Yuyuantan Conference on Research Integrity, hosted by the Institute of Scientific and Technical Information of China. The focus was on using artificial intelligence tools to assist, rather than distort, the research and publishing process. The conference released the Guideline on the Boundaries of AI-Generated Content Usage in Academic Publishing 3.0, providing detailed recommendations across various stages.
In the literature review phase, the guideline allows researchers to use AI-generated content (AIGC) tools to collect references on key terms or topics, categorize them, and summarize conclusions for background material. However, it warns that AI references may be fabricated or outdated, requiring manual verification of all citations and human-led decisions. For functions like 'deep thinking,' researchers must ensure credible source databases and cross-check recommendations to avoid AI hallucinations.
The guidelines extend to data collection, statistical analysis, figure generation, manuscript submission, peer review, and post-publication. Tan Tieniu, an academician of the Chinese Academy of Sciences and secretary of the Communist Party of China Nanjing University Committee, underscored researchers' primary responsibility as AI users and the need to standardize AI disclosure in papers. 'We should promote global collaborative governance, reach international consensus on AI usage norms and governance technologies, and advance Chinese standards,' Tan said.
Elsevier's recent Research of the Future report, based on a survey of over 3,200 researchers in 113 countries, found Chinese researchers are the most active AI users worldwide, with 69 percent adoption compared to the global average of 58 percent. About 66 percent of Chinese respondents believe AI will improve research quality. Gemma Hersh, Elsevier's senior vice-president and sales director for primary research, stated, 'China is always ahead when it comes to technology and innovation.' She added that China's research quality has significantly increased over the past five to 10 years, possibly aided by AI adoption.
Only 22 percent of global respondents consider AI trustworthy, while 45 percent lack sufficient AI skills. Hersh noted that enhancing transparency and traceability is a shared need, with Elsevier offering tools like Scopus AI, ScienceDirect AI, and the upcoming Leap Space. China's efforts in promoting compliant AI use are setting an international example. Geng Peng, an associate professor at Tianjin Foreign Studies University's School of Communication, said the guideline clarifies boundaries for academic publishing, research norms, and evaluation, helping researchers treat AIGC as a scientific tool and assistant with clear guidance.
The conference highlights AI's dual role in academia: accelerating innovation while necessitating strict norms to uphold integrity.