The use of AI in peer review or the integration of Artificial Intelligence (AI) is getting more and more noticeable in almost all academic studies, including the peer-review stage. Here are several ways AI is being utilized in peer review:
- Identifying Reviewers: AI algorithms can be used to find potential suitors for submitted manuscripts by looking at their area of expertise, publications and peer-review history. This will also help in deciding the reviewers and the experts so that the manuscripts are directed to the concerned persons of the field.
- Plagiarism Detection: Using AI-powered plagiarism detection software, peer reviewers can easily spot the cases of plagiarism or availability of similar work in the submitted manuscript. They use a search engine to compare the text of a document against the academic literature database to yield any possible cases of plagiarism.
- Language and Grammar Checking: AI tools with natural language processing (NLP) functionality can help peer reviewers spot such mistakes as grammar faults, spelling errors, and text phrasing lacks of harmony in the manuscripts. This also brings the readability of the manuscript into reviewing process.
- Quality Assessment: AI algorithms can help peer-reviewers in placing quality and rigor of research in presented manuscripts. That may mean understanding statistical methods, data analysis techniques, experimental design, compliance to reporting guidelines.
- Reviewer Suggestions for Editors: AI systems can analyze the content of the submitted papers and suggest the editor of the journal the names of possible reviewers, who are experts in the field in an appropriate way to evaluate the paper appropriately.
- Automated Reviewing: However, it is yet not widely adopted and in addition some researchers think about the way of using AI for automatization of some aspects of the peer review process. The automated processes could be useful in rating papers’ applicability of research results, or making first-stage screening of the papers, before the humans take part in the review.
- Reviewer Bias Detection: AI algorithms are able to make finding biases in peer reviewing by analyzing the comments of reviewers and their decision trends. Through the detection and correction of imbalances, journals can ensure more equal and just review mechanisms.
- Review Speed Enhancement: AI machines can help speed up the peer review process by executing governance-related activities, e.g. sending notifications to reviewers, following up the reviews review related performances, and performing communication functions between authors, reviewers, and editors.
Despite AI has the nature to improve the efficiency and quality of the peer review process, it is necessary to know its limitations and maintain the human oversight is a vital component to keep the peer review system’s integrity and fairness. The implementation of AI in peer review may have inherent risks of biases, privacy concerns, and lack of transparency. Thus, they should be addressed with considerable care.
Publishers Use AI in Peer Review
AI technologies are integrated into the peer-review processes of some publishers to enhance their processes in terms of the efficiency, accuracy, and also fairness.Here are some examples.
- Elsevier: Elsevier, one of the biggest academic publishers, uses AI at stages of the peer review. The AI-powered system “Editorial System Reviewer Explorer” by them evaluates reviewed papers for internal consistencies and preferences information to indicate the best reviewer for a submitted manuscript.
- Springer Nature: Springer Nature will be researching about the implementation of AI in peer review process to enhance the choice of reviewers, review of manuscripts and quality assessment. AI tools they develop use delved-in manuscript content, reviewer comments, and citation networks to suggest the reviewers and control manuscript quality.
- Frontiers : Frontiers, an open research publisher, uses AI algorithms to assist with text analyses, such as citation checks, grants review, and funding recommendations for manuscripts. Their tool “Artificial Intelligence Review Assistant (AIRA)” is AI-powered and assists the trained human reviewers in detecting potential problems in the submitted manuscripts e.g., plagiarism, data manipulation, and ethical challenges.
- Wiley: Wiley has incorporated AI technology into peer review, thereby helping in reviewer selection, plagiarism detection, and manuscript quality assessment The AI toolsreview manuscripts and use the data on manuscript content, reviewer expertise, and publication history to recommend capable researchers and streamline the publication process.
- Taylor & Francis: Taylor & Francis is looking into the application of AI in peer review with the intent of developing approaches to better match reviewers, improving manuscript review and determining quality assessments. Their AI-powered system, the `GRID database of digital science` , permits to identify candidates who could review with the right skills and helps analyze data from manuscripts easier.
- MDPI: The publisher MDPI, open-access, is using AI-based applications to filter out manuscripts and to assign suitable peer-reviewers. Their AI algorithms analyze manuscript content, authors’ style and reviewers’ experience to match the qualified reviewers and improve revision efficiency.
- Hindawi: Hindawi is a company that refines use of AI technologies to automate many steps of the peer review process, like reviewer selection, paper screening, and quality control. Such tools powered by artificial intelligence review manuscript contents, reviewer profiles, and citation networks to enable the right practices in the peer review process.
- IEEE: IEEE (Institute of Electrical and Electronics Engineers) utilizes AI tools into the peer review processes for journals and conferences publishings in engineering, technology, and computer science area. These AI systems help reviewer selection, detection of plagiarism, as well as evaluation of publication quality.
- Royal Society of Chemistry (RSC): RSC has been experimenting with Artificial Intelligence (AI) in peer review advisement on reviewer selection, manuscript scrutiny, and quality assessment for its scientific journals. The AI-based tools of theirs are designed to evaluate manuscript content, citation networks, reviewer experience, to increase the speed and accuracy of the revision process.
- Cambridge University Press: AI has been one of the focuses of Cambridge University Press and it has been used in AI technology in peer review processes in order to identify apt reviewers, screen manuscripts and as well as enhance quality assessment, respectively. The AI-based devices they create are specifically aimed at improving the efficiency and efficacy of the academic discipline-based peer review channel.
These are a few examples of those how the publishers employ AI technologies to enhance the different aspects of the peer review process. Incorporation of AI-based tools and algorithms by the publishers would enable the judicious use of the time, compliance with the principal of reviewer competence, upholding research integrity principles and augmenting the quality of the published scientific literature. Nevertheless, it’s very essential to keep on evaluating and rethinking these AI systems to resolve any shortcomings or biases, so as to protect the quality and also impartiality of the peer review process.
Reasons Publishers not using AI for Peer Review
Furthermore, the incorporation of AI in the peer review processes of the publishers differs among the publishers as some of them are not adopting it for several reasons. Here are a few possible reasons why a publisher might not be using AI for peer review:In the following are some probable motives why the publishing company is not making use of AI for peer review:
- Resource Constraints: The establishment, application, and sustainable of AI in peer review process may be a costly process that is both high in financial and human resources. Publishers, especially the small ones and those not capable of paying for AI facilities and/or experts, may not be able to afford the necessary investments in AI infrastructure and/or to find suitable AI experts to meet their needs.
- Technical Challenges: Therefore, the development of algorithms for quality assessment of reviewers and selection of reviewers will be some of the technical problems that should be faced while AI is introduced to peer review system. Finally, plagiarism detection should consider problems too. Publishers may be lacking the IT skills to build such tools that are smart and can be of great help in these situations.
- Risk Aversion: AI adoption to the peer review process comes at different levels of risk such as bias of algorithms in that process, data privacy and the reliability of AI systems. The truth could be that publishers might be cynical about the use of AI technologies without the enough proof of their validity and reliability.
- Preference for Traditional Processes: Additionally, publishers can prefer the traditional one in which the peer review process is editor based rather than the knowledge of people. Human reviewers may be more eager to have the request-personal approach and critical assessment features they offer, among other attributes that this may present, over the speed that AI-assisted systems may provide.
- Community Resistance: The academic publishing is conservative and does not want any changes, including those that concern fundamental processes such as peer review. Publishers may face resistance from authors, editors and reviewers who simply refuse to accept AI-powered peer review systems that might be new to them.
- Ethical Considerations: Publishers might have ethical questions about using AI in peer review, especially for topics like data privacy, algorithmic transparency, and potential bias in cases when an AI tool is autonomously making decisions.
Although the aforementioned points could clarify why some publishers do not use AI in peer review right at the moment, the fact of the adoption of AI in the academic world is horizontally striving. AI technologies are also evolving and their benefits becoming more apparent. This one reason that in the future more publishers may begin to experiment with and implement AI based systems for peer review.
Conclusion
Overall, the plunging of AI into the review mechanisms will definitely transform the scholarly publishing by increasing the efficiency, objectivity, and also quality. AI tools like NLP and machine learning models are the ways through which the researchers and publishers can automate their evaluation process, detect biases, and tighten the quality of their evaluations.
AI can help effectively in the various steps of the technical review including searching for the manuscripts, detecting plagiarism, recommending the reviewer, and even writing the review draft. AIs are now in a position to automate the routine tasks thereby enabling the reviewers to deliver the insights and recommendations relevant to the process that is quick and rigorous.
Moreover, AI-powered peer review frameworks can collect, interpret, and present a wealth of data, leading to a more detailed and refined evaluation of the research. Discernment and proficiency in picking up on the slightest variations and discrepancies that an average human reviewer may miss, makes their assessments more credible and also impartial.
On the other hand, it is very important to distinguish between AI as a complement which could find very useful application in the peer review and its replacement of the human expertize. The AI can speed up some parts of the process and also help the reviewer out but the human mind with it’s decision, contextual understanding and ethics will still be very much needed.
Moreover, issues such as the bias in the algorithms, privacy concerns about the data, and the requirement for monitoring, refining, and checking the performance of these systems will require a lot of attention to make the usage of these technologies ethical and fair in the peer review.
In brief, AI-assisted peer review can significantly revolutionize scholarship, bring in a lot of transparency, and also speed up scientific research. By integrating the capabilities of AI and the human reviewer together, we create a much higher quality, more efficient, and also inclusive peer review system that can extend beyond academia.
