Artificial intelligence (AI) is rapidly being integrated into all areas of life, including science and scientific publications. In this article, we will look at the opportunities AI opens up for researchers and the challenges they may face when using it to prepare scientific papers.
Advantages of using AI for writing scientific papers
- Searching for sources in seconds – the ability to quickly find the latest and most relevant scientific research to use in your work
- Improving text quality – checking for grammatical, stylistic, and punctuation errors, improving structure, selecting keywords
- Text translation – the ability to translate the text of a manuscript into any language in seconds
- Data visualization – creation of infographics, diagrams, visualizations, 3D models, and other graphic elements
- Processing large amounts of data – quick analysis of large amounts of information and acceleration of data processing
- Plagiarism check – identification of non-original fragments, including recognition of text generated by other AI.
- Search for new ideas – generation of new topics and possible directions for future research
Writing scientific papers using AI: problems
Violation of the principles of academic integrity
Since AI models are based on existing data, their incorrect use is considered a violation of ethical standards. Automatically generated texts have no independent scientific value and can be classified as plagiarism. AI can serve as a supporting tool, but it cannot replace the researcher's contribution.
Risk of obtaining false data
AI can generate false or fabricated data – for example, inventing a source of information or distorting statistics by referring to unverified data. This is especially dangerous in the scientific community, where accuracy and reliability play a key role. Information generated by AI must be verified.
Risk of false AI detection
Although scientific journals are implementing systems to detect AI content, these algorithms do not always work flawlessly. Sometimes texts written by humans are mistakenly identified as AI-generated. This highlights the limitations of existing tools and the need for a balanced approach to manuscript evaluation.
Superficial data analysis
AI is capable of quickly processing large amounts of information, but the depth of its analytics is limited. This is especially noticeable when outdated or irrelevant sources are used. High-quality academic research must be based on fresh, verified data from scientific articles, official statistics, and specialized databases.
Limitations of machine translation
Automatic translation is unable to convey terminological specifics and can distort the meaning of scientific texts. Poor-quality translation can even lead to an article being rejected by a scientific journal. That is why it is advisable to consult specialists when translating scientific manuscripts.
China's position on the use of AI in scientific works
In December 2023, the Chinese Ministry of Science and Technology published new “Guidelines for Responsible Research.” They impose obligations on scientific researchers, research institutions, research funding agencies, scientific and technological societies, and academic journals.
Provisions regarding the use of AI in scientific works:
- The use of content created with generative AI, especially content containing facts and opinions, must be clearly indicated. In addition, the process of its creation must be explained to ensure accuracy and respect for the intellectual property rights of others.
- As a rule, content marked by other authors as created using AI should not be cited as the original source. If citation is necessary, this should be explained.
- Unverified references created by generative AI should not be used directly.
- Generative AI should not be listed as a co-author of the results. The main methods and details of the use of generative AI should be disclosed in the relevant sections, such as Methods or Appendix.
- Authors should disclose information about the use of generative AI, specifying the specific name of the software, version, and time of use, and provide specific annotations to any supporting generated content that contains factual or subjective citations.
- When using generative AI in the peer review process, prior approval must be obtained from the peer review organizer (journal editorial board, conference program committee, etc.).
AI technologies have already been implemented in the largest international scientometric databases, such as Scopus and Web of Science. These technologies help scientists write scientific papers, and the number of benefits from their use continues to grow.
Therefore, AI is an effective tool that can improve the quality of scientific articles. However, it must be used in accordance with legal requirements and ethical principles, and the information generated by AI must always be verified. AI can help, but it will never replace scientists with their capabilities and contributions to the development of their field and science as a whole.
Scientific Publications is ready to help you at every stage of publishing your scientific work – from writing and selecting a scientific journal to successful publication and indexing. In addition, we provide our clients with post-publication support. Find out all the details during a free consultation with our manager – just fill out the form below or leave a request on the website. Together – to new achievements!
