Améliorer le discours académique à l’ère de l’intelligence artificielle: Une approche guidée pour une évaluation efficace par les pairs

ABSTRACT

The integrity of the peer-review process (PRP) is paramount in academic publishing and serves as a critical filter for scholarly output. This mini-review centers on the introduction of comprehensive guidelines, presented in tables format, aimed at streamlining the interactions between authors and reviewers during the PRP. These guidelines, derived from an in-depth exploration of the PRP, offer structured and practical advice to ensure constructive, transparent, and effective communication, especially related to the use of artificial intelligence. While this mini-review discusses the strengths and challenges of the current PRP, its primary focus is on providing tangible recommendations to enhance the quality and efficiency of the PRP. By providing explicit guidelines and emphasizing the cooperative essence of peer review, this mini-review aims to improve the PRP, ensuring that it remains a robust mechanism for upholding the highest standards of research and knowledge dissemination in an evolving academic setting.

introduction

The academic publishing landscape is a complex tapestry woven with rigorous standards, meticulous evaluations, and the collective pursuit of advancing knowledge (1-4).

Central to this elaborate system is the peer-review process (PRP), a mechanism that ensures the quality and integrity of scholarly outputs ( 4 - 7 ).

As the foundation of academic discourse, PRP is more than just a quality control measure; it shapes the trajectory of scientific dialogue and solidifies the credibility of research findings (4-7).

However, navigating this process, both as an author and reviewer, can be fraught with challenges (8- 13).

From aligning expertise to ensuring objective evaluations, the nuances of peer review demand clear guidelines and structured approaches (14 ).

Despite existing guidelines and/or recommendations (6 -20), pitfalls remain in PRP.

Reviewers may inadvertently introduce biases, overlook critical nuances owing to misalignment of expertise, or even fail to provide constructive feedback (6, 8, 17).

Authors, on the other hand, may misinterpret feedback, respond defensively to constructive criticism, or struggle to effectively communicate revisions (11-13, 15, 16)

Such challenges can lead to prolonged review cycles, potential misunderstandings, and even compromise the quality of the final publication (7 , 18-20 ).

Given the potential imperfections and drawbacks of the PRP, there is a pressing need for a comprehensive framework that addresses the unique challenges faced by both authors and reviewers ( 6 , 8 , 11 - 13 , 15 - 17 ).

Thus, this mini-review aimed to present a cohesive set of guidelines designed to streamline interactions and foster effective collaboration between authors and reviewers,ensuring transparent and constructive PRP.

The authors drew inspiration from their personal experiences acquired during their research activities, as potential readers for several international journals, and as members or leaders of editorial committees for certain international journals.

the importance of peer-review in academic settings

There are typically two primary categories of PRPs: open and blinded ( 6 ).

The open review system is characterized by complete transparency, where both authors and reviewers are aware of each other’s identities. In contrast, within the blinded review category, there are two subtypes: single-blind and double-blind reviews ( 6 ).

In a single-blind review, only the reviewers have access to the identities of the authors, and the authors themselves remain unaware of who is evaluating their work.

Conversely, in a double-blind review, both authors and reviewers were unaware of each other’s identity ( 6 ).

The peer review mechanism, foundational to scholarly publishing, is both a sentinel and catalyst in academia ( 5 , 14 ).

It not only ensures the quality and integrity of scholarly outputs, but also shapes and refines the trajectory of scientific discourse.

Exploring its multifaceted role, we recognize that peerreview is fundamentally collaborative, offering researchers invaluable external insights ( 21 )

This external perspective often illuminates areas ripe for refinement, ensuring research is robust and articulated in a manner resonating with its audience ( 7 , 18 - 20 ).

The iterative feedback loop refines the hypotheses, fortifies the methodologies, and crystallizes the narrative of the manuscript.

Beyond its role in refining research, PRP serves as a bulwark against academic malfeasances ( 22 ).

Meticulously examining manuscripts aids in unearthing instances of plagiarism, data falsification, and other breaches of scholarly ethics, ensuring that disseminated research is original and a credible addition to its field ( 6 , 8 , 17 , 22 ).

Moreover, the process plays a pivotal role in curbing undue self-citations ( 23 ), where authors might unduly reference prior work.

Through objective evaluation, reviewers ensure that citations are pertinent and justified, fostering a balanced academic discourse ( 6 , 8 , 17 ).

The advent of artificial intelligence (AI) in research presents a novel set of challenges ( 24 - 26 ).

Concerns have arisen regarding the undisclosed use of AI chatbots in research generation and/or assistance ( 24 - 26 ).

Such covert utilization is not only misleading but also breaches ethical standards ( 20 ).

Vigilant PRP can discern indications of machine-generated content, ensuring research authenticity and transparent AI involvement disclosure ( 26 ).

Additionally, peer-review champions ethical standards in research ( 20 ), and several international societies and publishers recommend peer-review ethics ( 4 , 27 ).

Beyond content evaluation, PRP ensures adherence to ethical norms in data acquisition, informed consent in human-centric studies, or humane treatment of animals in experimental research ( 6 , 8 , 17 , 20 ).

The expertise of reviewers is pivotal in pinpointing potential ethical oversights, ensuring that research aligns with overarching academic ethical tenets ( 6 , 8 , 17 ).

Although peer-review’s primary perception is a quality checkpoint, its role is expansive ( 6 , 8 , 17 ).

However, as academia evolves, the indispensability of peer-review to preserve the essence and credibility of scholarly publishing remains paramount ( 6 , 8 , 17 ).

Benefits of Peer-review for manuscript authors

Many authors frequently experience frustration and disappointment when faced with critical feedback ( 11 - 13 , 15 , 16 ).

Receiving comments from anonymous reviewers suggests that their paper can be intimidating ( 13 )

This emotional impact can be particularly pronounced when authors perceive feedback as uninformed, biased, or malicious ( 13 ).

Nevertheless, it is crucial to recognize that PRP plays a pivotal role in the refinement of manuscripts over time ( 13 ).

Experienced authors appreciate and leverage the benefits of constructive PRP ( 13 ).

Min ( 13 ) summarized the following six simple, yet important, tips to assist authors in responding appropriately to peerreviewers’ comments: i) Letter to the editor and reviewers, ii) Be polite and respectful, iii) Respond point-by-point to each and every comment raised by all reviewers, iv) Make the response self-contained, v) Stay optimistic; and vi) Check repeatedly for any mistake.

PRP offers manifold benefits to manuscript authors, serving as a bridge between the initial research conception and final publication ( 5 , 14 )

Peer-review provides authors with a unique opportunity to view their work through the lens of external experts ( 4 ).

This external evaluation often brings light nuances and perspectives that might remain obscured in the insular environment of individual research teams.

Such feedback can be instrumental in identifying gaps, suggesting additional analyses, or refining the research narrative to better resonate with the broader academic community ( 6 , 8 , 17 ).

Moreover, feedback from reviewers often encourages authors to think more critically about their work ( 6 , 8 , 17 ).

This prompts them to anticipate potential questions, address counterarguments, and present their findings in a more comprehensive and robust manner ( 6 , 8 , 17 ).

This iterative process not only strengthens the research but also enhances its accessibility and relevance to diverse readers.

Furthermore, PRP can serve as a valuable learning experience for authors, especially those in their early academic careers ( 6 , 8 , 17 ).

Engaging with positive and critical feedback fosters resilience, adaptability, and commitment to continuous improvement.

It also provides insights into the broader expectations and standards of the academic community, guiding authors in their future research.

Additionally, the validation received from a successful peer-review can bolster an author’s confidence in their research

This serves as an affirmation that their work is of value to the field and has the potential to contribute meaningfully to ongoing academic dialogues ( 6 , 8 , 17 ).

The demands of PRP are undeniably rigorous, yet the rewards for manuscript authors are substantial.

The end-result is a piece of work that stands not only as a testament to high-quality research, but also as a meaningful and relevant contribution to the broader academic discourse.

To assist authors in the process of responding to reviewers and/or the editorial team, we prepared response templates in the form of tables [Box 1 (English version), Box 2(French version), in the appendix].

Benefits of Peer- review for Peer-reviewers

Engaging in PRP positions reviewers at a pivotal juncture in academic advancement and professional development ( 6 ).

By undertaking this role, reviewers gain privileged access to emerging research, ensuring that they remain conversant with the latest developments in their disciplines ( 6 ).

This exposure not only facilitates continuous learning, but also provides insights that can be integrated into their own research endeavours ( 6 ).

The meticulous analysis required for PRP cultivates a heightened sense of critical evaluation ( 6 , 8 , 17 ).

Reviewers are tasked with assessing complicated research frameworks, a responsibility that enhances their ability to discern methodological nuances, recognize innovative approaches, and offer constructive feedback ( 6 ).

This continuous engagement with diverse research paradigms augments their academic wisdom and fosters a more comprehensive understanding of their fields ( 6 ).

Beyond their specific areas of expertise, reviewers benefit from exposure to a broad spectrum of academic perspectives ( 6 , 8 , 17 ).

This expanded purview encourages interdisciplinary engagement, promotes collaborative research endeavours, and provides a more holistic approach to academic inquiries ( 5 ).

For established academics, especially those in mentorship roles, insights derived from PRP are instrumental ( 6 )

They provide a comprehensive understanding of evolving research trends, prevalent challenges, and dynamic expectations of the academic community.

Equipped with this knowledge, they can offer more effective guidance to emerging researchers, ensuring their preparedness for the rigor of academic publishing ( 6 ).

From a professional standpoint,the act of reviewing is being increasingly recognized and documented (21).

Modern academic platforms such as Open Researcher and Contributor Identifier (ORCID) and Web of Sciences (WoS), which chronicle peerreview contributions, ensure that reviewers receive due acknowledgment of their efforts, enhancing their academic standing and contributing to their professional progression (21 , 28).

The role of peer-reviewers extends beyond service to the academic community ( 6 ).

It represents an opportunity for intellectual growth, professional recognition, and meaningful contribution to the refinement and advancement of scholarly discourse ( 6 ).

Ali and Djalilian( 6 ) have detailed general (n=16) and specific (n=25) good practices that should be followed to perform a good peer-review.

Pitfalls and challenges in the PRP

PRP, foundational to academic publishing, is not without its intricacies and potential pitfalls.

Both reviewers and authors, with their utmost dedication to scholarly rigor, can sometimes inadvertently encounter challenges that might affect the integrity and efficacy of the review.

While many scientists recognize the significance of PRP, a notable deterrent for many to participate as reviewers is the lack of adequate incentives.

For instance, journals such as MDPI offer open-access vouchers worth 100 CHF for reviewing manuscripts.

In many cases, this does not commensurate with the effort and time invested in the manuscript.

Furthermore, while journals such as Frontiers and MDPI disclose the names of reviewers in the published manuscripts, potentially enhancing the reviewer’s visibility and expanding their research network, this practice has not been universally adopted.

Other journals might merely offer modest discounts on open-access fees, typically ranging from 10 to 20%, or provide acknowledgment certificates.

While these gestures are commendable, they often fall short of reviewers’ expectations.

From our perspective, a more enticing approach would be to offer substantial open-access vouchers, for instance, 10% of the total open-access fees of the respective journal.

While this provides tangible incentives, it is crucial to ensure that such incentives do not make PRP appear lucrative.

The essence of peer-review lies in its commitment to scientific rigor, not financial gain.

By striking a balance, we can acknowledge the invaluable contribution of the reviewers without compromising the sanctity of PRP.

It is about fostering a culture of appreciation in the sciences, rather than veering towards a business-oriented model.

A salient challenge for reviewers is the occasional misalignment of their expertise.

Despite editors’ diligent efforts to align manuscripts with reviewers of pertinent expertise, there can be instances where a reviewer is tasked with evaluating a manuscript that is tangential to their primary domain.

Such situations can lead to feedback that may miss certain specialized nuances, or in some cases, result in overly critical evaluations stemming from varying academic viewpoints.

Bias, both conscious and unconscious, remains a significant concern in the PRP.

Reviewers, being human, might unintentionally allow personal biases to colour their evaluations.

These biases can arise from myriad sources such as prior academic interactions, institutional affiliations, or even differing academic philosophies, potentially affecting the objectivity of the review.

Deciphering and navigating feedback from reviewers, especially when faced with divergent critiques, can be a complex task.

This necessitates a fine balance between addressing reviewers’ concerns and maintaining the core essence of the research.

Moreover, the challenge of differentiating between constructive feedback and comments that might be unduly harsh or slightly offmark often arises. The potential conflicts of interest are also large.

There are instances where authors might recommend reviewers, with whom they share professional or personal ties, hoping for a more favourable review

However, reviewers might have undisclosed professional histories with authors, which could inadvertently influence their evaluations.

Furthermore, many of the imperfections or pitfalls that arise during PRP can be attributed to miscommunications or tasks executed improperly by either side.

Such missteps can lead to inconsistent responses from reviewers, increasing the likelihood of manuscript rejection.

Conversely, this can result in reviewers providing feedback that does not constructively contribute to the improvement of the manuscript.

Recognizing these challenges and the pivotal role of effective communication and proper task execution in PRP, we feel the need to provide detailed guidelines.

These guidelines aim to prevent such missteps, aid reviewers in constructing organized and constructive feedback (Table 1), and ensure more consistent responses from the authors (Table 2).

By doing so, we hope to reduce the rejection rates and enhance the overall quality of the manuscript.

To this end, we delineated the clear guidelines in Tables 1 and 2.

These tables serve as a comprehensive set of guidelines designed to mitigate the aforementioned challenges and ensure that PRP remains transparent, objective, and constructive, fortifying the foundations of academic discourse

Quoting a fascinating editorial published in Nature ( 29 ) “" For authors: in the interests of robustness and genuine impact, resist the pressure to publish prematurely.For referees: please do not ignore any impulse to demand more, but be self-critical too.”

Position of the international committee of Medical Journal Editors(ICMJE)(4) regarding AI assisted technology

According to the ICMJE ( 4 ), upon submission, it is essential for the journal to mandate authors to disclose whether they employed AI-augmented technologies, such as large language models, chatbots, or image generators, during the creation of their submitted work.

ICMJE recommends that authors who utilized such technology must provide a detailed account of its application in both the cover letter and the submitted work ( 4 ).

The ICMJE noted that chatbots, chatbots like generative pre-trained transformer, should not be credited as authors, as they lack the capability to assume responsibility for the accuracy, integrity, and originality of the content, which are prerequisites for authorship ( 4 ).

For the ICMJE, the responsibility for any submitted material that incorporates AI-augmented technologies lies squarely with human authors ( 4 ).

According to the ICMJE, authors must exercise careful scrutiny and editing of the output generated by AI, as AI has the capacity to produce content that sounds authoritative, but may be erroneous, incomplete, or biased ( 4 ).

It is imperative that authors do not attribute authorship or co-authorship to AI or AIaugmented technologies, nor should AI be cited as an author ( 4 ).

Furthermore, authors must affirm the absence of plagiarism in their paper, encompassing both text and images generated by AI ( 4 ).

Humans bear the responsibility for ensuring that proper attribution is given to all quoted material, complete with comprehensive citations ( 4 ).

Peer-review and the future of publishing : is it time to transition from the human-dependent Peer-review system to a completely AI-driven one?

In his reflective essay titled «Ending human-dependent peer review» published on September 29, 2023, Irfanullah advocates for a comprehensive shift away from the human-dependent peer-review system in favour of a fully AI-based approach (30).

Irfanullah provides seven compelling reasons to support this proposal (30).

First, he underscores the current system’s inequity, where voluntary peer-review often goes unacknowledged as an academic responsibility by universities ( 30 ).

While some institutions recognize this service, many do not, placing an unfair burden on select academics ( 30 ).

Second, he questions the necessity of this system to keep peer-reviewers informed about new research, noting the abundance of readily available information sources ( 30 ).

Third, Irfanullah delves into the concept of «good karma» in peer-review, where reviewers assist others in their field, expecting reciprocal assistance ( 30 ).

However, this reciprocity does not always materialize, leading to frustration ( 30 ).

Fourth, he highlights concerns regarding the exploitation of peer-reviewers, particularly in the context of reviewing research project proposals ( 30 ).

Fifth, he addresses the issue of peer-review ghost-writing and suggests that the concept of «co-reviewers» might exacerbate existing inequities ( 30 ).

Sixth, he explores the significant amount of time that reviewers invest in the PRP without tangible benefits, despite recent efforts to recognize their contributions (30).

Seventh, Irfanullah criticizes the high costs associated with publishing and suggests that the current system primarily benefits publishers at the expense of reviewers (30).

Ultimately, Irfanullah proposes a five-phase transition towards a fully AI-based review-system, emphasizing the necessity of adequate AI training to mitigate algorithmic limitations (Table 3) (30).

conclusion

PRP remains a cornerstone in the edifice of academic publishing, acting as a rigorous filter that ensures the dissemination of research that meets the highest standards of scholarly excellence.

This investigation provided a comprehensive examination of the peerreview mechanism, emphasizing its pivotal role in shaping the trajectory of scientific dialogue, refining research contributions, and safeguarding the credibility of academic literature ( 31 ).

Concurrently, we identified and discussed the potential challenges and pitfalls inherent to PRP, underscoring the necessity for effective communication and adherence to established best practices.

A significant outcome of our mini-review was the formulation of detailed guidelines tailored to both the authors and reviewers.

These guidelines are envisaged to optimize PRP, engendering a more collaborative and constructive milieu.

These guidelines offer a roadmap for navigating reviewer feedback, enhancing the robustness and relevance of their manuscripts.

For reviewers, the guidelines present a structured framework, ensuring that the evaluations are thorough, objective, and conducive to improving the manuscript.

The broader implications of this mini-review extend beyond the confines of academic publication.

By facilitating a more streamlined PRP, we anticipate a series of cascading benefits across scholarly ecosystems.

Improved manuscript quality translates to research that is not only academically rigorous, but also possesses the potential to influence policy decisions, inform industry standards, and enhance public comprehension of intricate subjects

Moreover, by mitigating the challenges in PRP, the research community stands to benefit from a more expedient dissemination of knowledge, catalyzing the pace of academic advancement ( 32 ).

Finally, after one year of use these guidelines, it is recommended to assess their usefulness by conducting an online survey among our reviewers, and to publish the results in Tunis Med.

APPENDIX: Tunis Med: Templates to assist authors in the process of responding to reviewers and/or the editorial team (English (Box 1) and French (Box 2) versions) available via this URL: https://doi.org/10.5281/ zenodo.10030627.

DECLARATION. In order to correct and improve the academic writing of our paper, we have used the language model ChatGPT.