(The opinions expressed in this article are those of the author and do not necessarily reflect the opinions of Al Fana's media).
Is it okay to use AI in scientific research? A recent study by the Nature Journal shows that the question has been featured in more than 5,000 academics around the world, and that scientists are deeply divided on the ethically acceptable use of artificial intelligence.
The findings explained in the article entitled “Is it OK for AI to write scientific papers?” highlight a complex and often contradictory landscape. The disparities among researchers go beyond mere differences and point to deep concerns about the future of scientific research and the ethical principles that underpin it.
There is a growing trend towards viewing AI as a promising tool to accelerate the process of writing scientific papers, especially in the early stages of manuscript preparation. Large-scale language models (LLMS) can help English-speaking researchers overcome language barriers, summarise a wide range of literature, propose logical article structures, and allow researchers to concentrate more on scientific content and in-depth analysis.
AI Ethics in Scientific Research
However, despite this possibility, there is growing concern that artificial intelligence can undermine core principles of scientific research, especially in areas such as content liability, originality and the risk of unintended plagiarism. Particularly given the tendency to generate false or manufactured information known as “hatsui,” pressing questions emerge, especially among those with legal and ethical accountability, about errors or misleading content generated by AI.
Furthermore, AI models are trained on vast datasets, which can inherit and persist existing biases, leading to distorted and stereotyped output. This poses a serious threat to the objectivity that scientific research requires.
Also, many scholars fear that excessive reliance on AI tools can undermine researchers' ability to think critically and reduce their ability to express themselves.
Most concerning is the gap highlighted by natural research between the broad theoretical acceptance of AI as a research aid and the prominent reluctance to apply it in practice. What's even more troubling is the fact that a significant proportion of researchers using AI tools choose not to disclose this in their submitted papers.
This action raises serious concerns about academic transparency and integrity, and raises questions about the scientific community's ability to trust research that could pose traces of AI without proper awareness. Such unwillingness can be attributed to the fear of stigmatization or the lack of clear, standardized guidelines for when and how to report AI use in academic work.
The urgent need for guidelines
The Nature Survey findings highlight the need for immediate and decisive action by academic institutions, publishers, and research institutions to establish a clear ethical and legal framework that governs the use of AI in scientific writing. Large publishers like Elsevier and Springer have already begun implementing policies. Many of them prohibit the listing as AI co-authors, but allow it to be used as a tool, as long as there is explicit disclosure in the methodology or acknowledgements.
Key issues include defining authors when AI plays a critical role in the writing process, ensuring unbiased content embedded in training data, and protecting researchers' ability to think critically and express themselves. These are not minor considerations. They are the basis of the reliability of science and its role in advancement of knowledge and in the service of society.
However, it is important to recognize that inherent limitations of research variables can lead to unintentional bias in data, particularly medical and biological research bias. These limitations are often difficult to address for technical reasons. As a result, the model can suffer from problems such as overfitting and poor generalization. There, it works well with training data, but when applied to new, invisible data, it is not possible to accurately predict the outcome.
It is noteworthy that the discussion of data bias is often focused on intentional bias (generally related to the social and humanities), but unintended biases derived from methodological or technical constraints are more typical in scientific research.
The future of scientific writing
Natural research is undoubtedly more than a simple opinion. It presents a pivotal moment in the ongoing debate about the future of scientific research in the age of artificial intelligence. It is forced to deeply reflect and act proactively in the best ways to integrate these powerful, transformative technologies into the research ecosystem without compromising long-defined core values, such as long-defined scientific pursuits, the essential role of rigorous accuracy, objective analysis, intellectual originality, and human creativity.
This shift requires more than a simple policy revision. The concept of “author” must be comprehensively and intentionally redefine. It covers meaningful intellectual contributions and full accountability to both content and its potential outcomes, not merely an act of writing or drafting.
To maintain the integrity of scientific research, it is essential to establish clear and enforceable guidelines, foster a culture of full transparency, and request explicit disclosure of AI support at either stage of the research or writing process.
Equally important is the development of sophisticated and reliable tools that can detect AI-generated or modified content. These measures are important to ensure that science remains a reliable, ethically-based source of current and future human knowledge.
Furthermore, this study highlights the urgent need for continuous and constructive dialogue between all stakeholders, including researchers, policymakers, AI developers and ethicists, ensuring that technological advancements are consistent with the noble purpose of science and human advancements.
Such ongoing conversations are essential to addressing unexpected challenges and maintaining the mind of curiosity and critical thinking at the heart of scientific innovation. Rather than simply adapting to this new era, this challenge presents a valuable opportunity to critically reassess and update scientific practices to suit its demands.
Mohamed Al-Rubei is a professor emeritus and Conway Fellow at Dublin University. He specializes in biology and focuses on science and technology in the Arab world.
