Can you use generative AI in peer review process?

Discover why generative AI tools cannot be used in manuscript review and learn about alternative technologies to enhance this process.

Can you use generative AI in peer review process?

As generative AI continues to make inroads into academia, reviewers are asking whether this technology can be used in the peer review process of journals. 

To answer this question, let’s take a look at the policies of some major publishers. But first, it’s important to remind ourselves why peer reviewers are so crucial to academic publishing. 

Peer review “functions to encourage authors to meet the accepted high standards of their discipline and to control the dissemination of research data to ensure that unwarranted claims, unacceptable interpretations or personal views are not published without prior expert review,” reads a paper entitled “Peer review – Why, when and how.” 

When it comes to generative AI, many publishers require reviewers not to use this technology in the review process, primarily due to concerns about confidentiality, bias, and accuracy of information. This applies both to the process of reviewing the manuscript and to the writing of the review report

Elsevier argues, for example, that reviewing a scientific manuscript involves responsibilities, such as critical thinking and original assessment, that only humans can perform. It cautions that generative AI can lead to incomplete, incorrect, or biased conclusions about the manuscript. 

With regard to confidentiality, Elsevier states that when a researcher is invited to review the work of another researcher, “the manuscript must be treated as a confidential document.” 

It asks reviewers not to put a submitted manuscript, or any portion of it, into a generative AI tool, as “this may violate the authors’ confidentiality and proprietary rights and, where the paper contains personally identifiable information, may breach data privacy rights.” This confidentiality requirement also applies to the peer review report.

In a similar vein, Springer emphasises the limitations of generative AI tools, noting that they “can lack up-to-date knowledge and may produce nonsensical, biased or false information.”

“Manuscripts may also include sensitive or proprietary information,” it adds, asking peer reviewers not to upload manuscripts into tools powered by generative AI.

These publications, along with many others, hold reviewers accountable for the accuracy of the information and views expressed in their reports, highlighting that reviewers must meet the highest standards of integrity in this process.

See also: Why InstaText is your best bet for getting published 

Alternative technologies

Now the question is: Are there other technologies that reviewers can use to improve the language and readability of the texts they share with authors, editors, or co-reviewers, while avoiding the confidentiality, hallucination, and bias concerns associated with generative AI? The answer lies in tools powered by advanced language technologies, but not by generative AI.

These tools—which include editing assistants, grammar checkers, and proofreaders—only analyse and improve the existing text instead of generating new content or inserting new information. This ensures that the intended message is communicated clearly, reducing the risk of misinterpretation and enhancing the efficiency and effectiveness of the review process.

InstaText is one such tool. As an advanced editing assistant, it does not use generative AI, i.e. it does not generate new content, but only improves the existing text in terms of aspects such as readability, clarity, and grammar.

InstaText Editor

It also has an interactive editing style that gives users full control over the editing process, including the option to accept or reject suggested changes. InstaText also never stores texts

If you’re already a user, you can read our blog posts to get information and insights that can help you in your academic endeavors. If you’re not a user yet, you can try InstaText for free and see for yourself what it’s like to work with InstaText. 

Comparison: ChatGPT vs. InstaText vs. Grammarly

Disclaimer: Although InstaText has been designed to meet rigorous academic requirements, it is important to check each publication’s specific guidelines regarding the use of editing tools. Additionally, while InstaText never stores texts, other tools using similar technologies might. Always read the privacy policy of each tool carefully before using it.

“This tool is outstanding, exceeded my expectations. I’m used to using Grammarly but InstaText is a more thorough tool and comes up with much better suggestions for rewrites. A game changer for editing.”

— Stephan Skovlund, Business Consultant

“InstaText makes your text engaging to read, coherent, and professional-looking. Further, I feel that paragraphs corrected by InstaText look akin to what I see in top marketing and social psychology journals. It is a huge help for an academic writer because rather than focusing on making the text appealing, you can simply focus on what you want to say and build a logically unfolding narration.”

— Dr. Michał Folwarczny, Postdoctoral Researcher

“It’s like having a trusty editor by my side 24/7! ✍️💙 I’ve experimented with different text editing apps over the years, but none of them come close to InstaText. It helps me get my point across super clearly and efficiently while freeing up time for refining and developing my ideas. I’m very glad to have a “second set of eyes” that can quickly spot errors, inconsistencies, or confusing phrasing I might have missed, saving me time and effort in editing. As someone who writes a lot, InstaText has become a go-to editing tool for me. I believe it should be in every writer’s toolkit!”

— Saša Tatarevič, Digital Product Designer