What you need to know about artificial intelligence in research and publishing

Reading time
7 mins
What you need to know about artificial intelligence in research and publishing

Artificial intelligence (AI) makes it possible for machines to “learn” and perform human-like tasks. Every industry is now implementing some form of AI, and these applications are continuing to grow. AI has become a part of our daily lives too. Moments ago, as I sat typing this, AlexaTM informed me that based on my previous purchases, it might be time to re-order microwaveable popcorn!

As with nearly any industry, AI technology has made inroads into various processes in the publishing industry. The scope and applications of AI throughout the publication process are tremendous and constantly evolving.

AI can assist authors, editors, and publishers

AI in publishing is no longer just a novelty; it is being applied at various points in the publication pipeline. Further, the growing volume of open access scholarly content, including datasets and code, provides a rich resource for training datasets.

Authors can harness AI tools to speed up the publication cycle

Several AI offerings can support different steps in the scholarly publication cycle. AI-powered tools for literature discovery and summarization can free up time for other research activities. AI-trained platforms can assist with grammar and language checks and formatting checks. Further, AI-based pre-peer review screening can guide authors to revise their paper before it is sent for peer review. This reduces the chances of desk rejection and can shorten the peer review period.

Editors and reviewers can benefit from AI tools for faster turnarounds

At journal editorial offices, AI tools have the potential to take the load off various tedious tasks—managing astronomical submission volumes, increasing process efficiency, and developing more efficient peer review processes. AI has been used to assist with selecting journals, identifying a paper’s subject matter, determining if the subject falls within the journal’s scope, suggesting reviewers, assessing language quality, detecting plagiarism and duplicate submission, formatting documents, and assessing the appropriateness of experimental design and statistical analyses.1 Other opportunities include assessing the novelty of a study and checking for ethical compliance, copyright issues, and image duplication. The opportunities for AI in publishing are expected to keep growing and evolving rapidly.

AI-related developments in the industry

Many AI tools use machine learning and natural language processing of big datasets. Generative Pretrained Transformer 3 (GPT-3), created by OpenAI, is a language model that generates fluent strings of text after deep learning from a mind-boggling number of books, articles, and websites.2

Let’s take a look at some important advances in AI in scholarly publishing.

1. Article summarizer tools

In this era of information overload, it is impossible to read all the content pertinent to a topic. TL;DR (too long; didn’t read)—a relatively new acronym in daily parlance—aptly captures our daily struggle with information overload. At some point, we have all wished for a long text to magically get converted to a concise, readable extract. TLDR This does exactly this, i.e., “human-like summarization.” This free tool uses advanced AI models to generate the gist of any text quickly, without the user having to read all the paragraphs. Other text summarizing tools are Scholarcy and UNSILO (a Cactus Communications brand), which automatically extract key concepts to summarize manuscript content.3 Summarizer tools can be used at the pre-peer review stage to check subject matter and relevance to a journal’s scope. Such tools can also be of great help to an author at the literature review stage.

2. Literature discovery tools

Machine learning can help develop algorithms that follow what you read and why. This can translate into a great time-saving tactic for researchers who have reams of literature to sift through. R Discovery is a free literature discovery app; once you set up your areas of interest, the app finds the top three papers and presents them in the form of a daily feed. R Discovery is AI driven, offering customized research reading by learning the user’s reading interests. It also intuitively provides key highlights, summaries, and roundups of research relevant to the user’s field.

3. Identification of image manipulation

Computer vision is a field of AI in which computers are trained to interpret and understand visual elements. Using digital images and deep learning models, machines can accurately identify and classify objects. Publishers are beginning to adopt AI to spot image fraud in submitted papers. Early last year, journals published by the American Association for Cancer Research (AACR) began to carry out such added checks on manuscripts provisionally accepted after peer review. They use AI-powered software from Proofig, which alerts editors to duplicated images, as well as those with parts that have been doctored by rotation, flipping, or stretching.4

4. AI-powered academic editing

There are numerous language-support services available to researchers today. Many researchers may need quick checks of their manuscripts if they do not have the time or the skill level needed to have them checked for language. AI tools can come in handy in such cases. An example of such a tool is Paperpal, an AI-powered language-correction service offered by Cactus Communications. This tool has been trained on millions of research papers edited by expert editors from various disciplines. Besides effectively correcting language errors, it also ensures the correct use of domain-specific terminology.

Human oversight is needed with AI in research and publishing

The advancements in AI in research and publishing are exciting. However, complete reliance on AI for any of the mentioned steps should be avoided. For instance, editorial decisions indicated by AI must involve an editor on the acceptance or rejection of a manuscript. Similarly, in flagging research misconduct based on AI, human intervention for verification and final decision-making is essential. Further, AI tools are prone to biases that exist in the databases they are trained on; any potential biases should be corrected and the tools updated accordingly.

Another unwanted outcome of complete dependence on software screening is that some authors try to use AI to override the software, e.g., tweaking text to escape plagiarism detection (case in point: tortured phrases). Sophisticated AI-powered generation techniques can produce text indistinguishable from that of humans. A team of computer scientists led by Guillaume Cabanac identified scientific texts with tortured phrases or “unexpected weird phrases in lieu of established ones” in reputable journals.5 (A fitting example of a tortured phrase is “counterfeit consciousness” instead of “artificial intelligence”!) This discovery by Cabanac et al. only underscores the irreplaceable human role in AI technology, and that AI in publishing should serve as an alerting mechanism and support stakeholders in making informed choices.

Wrapping it up

Opportunities for AI in publishing are expanding rapidly. AI provides exciting opportunities for designing intelligent products and devising novel service offerings. Aided by an array of AI solutions, authors, editors, and publishers can perform their tasks with increased efficiency.


1. COPE Council (2021). COPE Discussion Document: Artificial intelligence (AI) in decision-making. https://doi.org/10.24318/9kvAgrnJ

2. Hutson, M. Robo-writers: the rise and risks of language-generating AI. Nature 591, 22–25 (2021). doi: https://doi.org/10.1038/d41586-021-00530-0

3. Checco, A., Bracciale, L., Loreti, P. et al. AI-assisted peer review. Humanit Soc Sci Commun 8, 25 (2021). https://doi.org/10.1057/s41599-020-00703-8
4. Van Noorden, R. Journals adopt AI to spot duplicated images in manuscripts Nature 601, 14–15 (2021). doi: https://doi.org/10.1038/d41586-021-03807-6

5. Else, H. (2021) ‘Tortured phrases’ give away fabricated research papers. Nature 596, 328–329 (2021) doi: https://doi.org/10.1038/d41586-021-02134-0

Be the first to clap

for this article

Published on: Feb 18, 2022

Sunaina did her masters and doctorate in plant genetic resources, specializing in the use of molecular markers for genotyping horticultural cultivars
See more from Sunaina Singh


You're looking to give wings to your academic career and publication journey. We like that!

Why don't we give you complete access! Create a free account and get unlimited access to all resources & a vibrant researcher community.

One click sign-in with your social accounts

1536 visitors saw this today and 1210 signed up.