The use of Artificial Intelligence (AI) assisted technologies in scholarly publishing has come a long way over the past two decades. From simple word processing tools to more sophisticated applications, AI is being used extensively to optimize work processes for both authors and publishers alike. By taking on tedious tasks like proofreading, editing, translations, detecting plagiarism, and identifying data misinterpretation, AI in research is helping to speed up and ease the publication journey and free up valuable researcher time.
Use of AI by scientific and academic journals
A recent Dimensions database study has revealed that global research output increased exponentially over the past few years with 4.7 million articles published in 20201. Academic journal editors and peer reviewers have been struggling to stay afloat given this deluge of scientific literature being submitted for publication. It has been estimated that over 15 million hours are spent every year just on reviewing previously rejected manuscripts, which are resubmitted to other journals2. Today, AI is being used to help make things easier and quicker for those in the scientific publishing arena.
- Identifying and assigning peer reviewers: The use of AI tools in research publishing is helping editors with the increasingly difficult task of identifying peer reviewers in specific research disciplines. Editors can now avail an unbiased ranked list of the most suitable peer reviewers based on the manuscript subject in just a few seconds – a task that would have otherwise taken hours to complete.
- Summarizing manuscripts: AI-powered tools are now being used by authors and journals to summarize submitted manuscripts and highlight important data on the study, participants, methods used, ethical compliance information, and results. This gives a better overview of a manuscript than the keywords typically submitted by authors and makes it easier for editors to identify papers that can be taken forward for publishing.
- Checking for publishing readiness: AI in research is also being used extensively to improve author submissions and optimize editor workflow. Today, there are a number of AI tools that can help to rate a manuscript’s language quality and provide suggestions on how to edit and polish it further. Journal editors can use such AI tools to assess whether the manuscript meets the basic journal requirements, including having the right style, reference format, and article structure. This allows editors and reviewers to focus on the content of the paper instead of the tedious task of proofreading.
- Assessing consistency of statistics: Today, artificial intelligence research tools like Statcheck, developed by Michèle Nuijten, a methodologist at Tilburg University in the Netherlands and colleagues, helps extract and check statistics from psychology research papers. Interestingly, Nuijten and his team found that about 50% of the papers that were published in psychology journals had at least one statistical error and in fact, in one out of every eight studies, the inaccuracy was significant enough to change the statistical significance of the published result.3
Use of AI tools by authors and researchers
Artificial intelligence is easing the work of over-burdened authors and researchers. Instead of having to spend countless hours tediously sifting through articles, proofreading, editing, checking for inconsistencies, and identifying the right journals for submission, researchers are now using AI-based algorithms and tools to increase efficiencies.
- Help with journal selection: For most researchers, the decision of which journal to submit their manuscript to is often a daunting and confusing one. There are numerous things to think about when choosing the right journal, including the quality of the journal, the length of time it takes to publish an article, and if the research meets the journal’s purpose and scope. Today, AI tools are making it easier for authors to find an ideal journal match with the aid of artificial intelligence techniques.
- Easing the research and literature-discovery process: The process of sifting through stacks of journals, both online and offline in search of relevant literature can be tedious and time consuming. Artificial intelligence for publishers and researchers is making this easier. Literature search can be personalized by taking into consideration specific subject areas and keywords. A good example of this is the R Discovery app, which uses your reading preferences to sift through a large repository of scientific literature and recommend the most recent and relevant papers for researchers.
- Tools for identifying plagiarism: Researchers are increasingly using AI software to help them detect possible duplication and unintended plagiarism so that they can take corrective measures before submitting their paper to a journal. There are AI-based tools that use deep learning, neural networks, machine learning, as well as natural language processing to evaluate your article and produce accurate results that can help avoid plagiarism.
- Quicker, language-editing and proofreading: AI tools are increasingly being used for tasks that humans have traditionally carried out, at a greater scale and speed. One such rather tedious task is proofreading and editing. Most researchers develop a blind eye to language or grammatical errors because of the many times they read and re-read their manuscripts. For researchers unfamiliar with English, the process becomes even more cumbersome. AI-based language-support tools such as Paperpal are now considered indispensable by researchers who seek to submit an error-free manuscript to journals for publication. Such tools offer authors substantial advantages and are very useful in not only identifying grammatical errors but also in ensuring accurate and consistent use of technical terminology, which enhances the overall quality of manuscripts.
Given its many advantages for researchers and journal editors alike, it is clear that the use of AI in research is only set to increase. With new developments and improvements in technology, AI-based tools will not only help make the research landscape more efficient, it will also help save precious time and consequently reduce costs in the existing publication process – perhaps even transforming it completely. However, although AI will be adopted more widely, human judgement in the creation and evaluation of research papers will remain a critical factor. The fundamental reason behind this is that AI tools can think like humans, but cannot comprehend those thoughts. Hence, a blend of human intervention with AI in research will remain essential for success.
References
- Burley, R. How AI is accelerating research publishing. Research Information, January 2022. Available at https://www.researchinformation.info/analysis-opinion/how-ai-accelerating-research-publishing
- Kelly, J, Sadeghieh, T, Adeli, K. Peer Review in Scientific Publications: Benefits, Critiques, & A Survival Guide. PubMed Central, October 2014. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4975196/
- Baker, M. Smart software spots statistical errors in psychology papers. Nature, October 2015. https://www.nature.com/articles/nature.2015.18657
