Use of AI in Scholarly Publishing: Striking the Balance Between Innovation and Integrity

AI is taking an increasingly prominent role in scholarly publishing, offering the promise of streamlining editorial processes, improving efficiency, and reducing human error. However, as we embrace AI’s potential, the critical question remains: Can we use AI without compromising the core values of research integrity? Are we prepared to trust AI’s judgments over human expertise? Let’s take a closer look at how AI can be ethically applied in scholarly publishing—enhancing rather than overshadowing human contributions.

1. Pre-Editing: Assisting, Not Replacing Authorial Voice

AI-driven language editing tools can be invaluable, particularly for authors whose first language is not English. These tools can help refine grammar and clarity, ensuring that the author’s ideas are conveyed effectively. However, there is a risk of allowing AI to replace the authentic voice of the author rather than simply enhancing it. AI should be seen as a tool for improvement, not a substitute for the author’s unique perspective. The rise of AI-generated articles raises ethical concerns, blurring the line between enhancement and authorship. With the increasing threat of “papermills”—which sell fake authorship and fabricated research—AI could be exploited to produce work that lacks originality. The challenge is to maintain the human essence of scholarly writing while leveraging AI for efficiency.

2. Technical Assessment: Complementing Human Review, Not Replacing It

AI can significantly aid the technical assessment of manuscripts by quickly identifying missing metadata, formatting issues, or incomplete sections. This can reduce delays in the editorial process and ensure a smoother transition to peer review. However, while AI can streamline these preliminary checks, it cannot replace human review entirely. Human editors must still ensure that editorial standards are met and that no critical nuances are overlooked. AI should assist, not replace, the essential role of human oversight in maintaining quality and consistency.

3. Editorial Checks: Supporting Plagiarism and Image Manipulation Detection

AI tools that detect plagiarism, self-citations, or image manipulation are becoming a cornerstone of editorial checks. These tools are crucial for maintaining the integrity of research, but they raise an important question: Can we rely on AI alone for these sensitive checks, or should human expertise still play a central role? AI can alert editors to potential ethical violations, but human judgment is essential to interpret context, intent, and the subtleties of a case. Experienced editors should be the ones to make final decisions, ensuring fairness and accuracy in handling these red flags.

4. Reviewer Selection: Enhancing Diversity While Maintaining Transparency

AI can help identify qualified and diverse reviewers by analyzing manuscripts and generating relevant keywords. This approach broadens the pool of potential reviewers and can reduce biases that are common in traditional reviewer selection methods. While this is a step forward, it’s important that AI is not relied upon to make reviewer selections autonomously. Human editors must retain the final say to ensure transparency and prevent over-reliance on automated decisions. The goal should be to use AI to complement human judgment, not replace it.

5. Research Evaluation: AI as a Supplement, Not a Substitute

AI can assist in the peer review process by providing preliminary evaluations or flagging areas that require closer attention. While this can streamline the process, AI can never fully replace the nuanced understanding and expertise that human reviewers provide. The role of AI should be to support, not supplant, the critical analysis that human reviewers contribute. Similarly, using AI to generate responses to reviewer comments may seem tempting, but this risks losing the depth and insight that only human experts can provide. Human judgment should always remain at the core of manuscript evaluation.

6. User Identification: Ensuring Integrity and Transparency

User identification is a critical aspect of maintaining research integrity. AI can play a vital role in detecting duplicate accounts, spoofed emails, and undisclosed conflicts of interest between authors and reviewers. However, when using AI for identity verification, we must ensure that fairness, privacy, and transparency are upheld. It’s essential that AI is used to support—rather than replace—traditional identity checks, ensuring a robust and transparent verification process.

7. AI for Data Integrity: Verifying, Not Replacing Human Expertise

AI shows great potential in verifying the integrity of data within research articles, especially when working with large datasets. By identifying outliers, inconsistencies, or errors in data visualization, AI can assist editors in maintaining accuracy. However, AI should never be the sole arbiter of data accuracy. While it can act as a valuable safety net, human experts must retain oversight to interpret the data’s context and ensure its validity. AI’s role here is to complement human expertise, not replace it.

8. Automating Metadata and Reference Management: Time-Saving, But with Caution

Metadata and reference management are crucial for indexing and discovery, but they are time-consuming tasks. AI can automate much of this work, making it easier for authors and editors to format references and ensure proper indexing. However, relying on AI to manage metadata and references carries the risk of errors. A poorly formatted reference or inaccurate metadata can impact an article’s accessibility. Therefore, while AI can save time, editorial oversight remains necessary to ensure accuracy and consistency.

9. Encouraging Transparency in AI Usage

Transparency in how AI is used in scholarly publishing is essential for maintaining trust. Journals should be open about the role AI plays in manuscript evaluation, ensuring that stakeholders understand its limitations and contributions. Clear guidelines on AI usage should be established, so that authors, editors, and reviewers are aware of the tools in use and how they affect the publishing process. This openness will help manage expectations and ensure that AI is used ethically and responsibly.

Final Thoughts

AI has the potential to significantly enhance research integrity in scholarly publishing, but only if it is used thoughtfully and ethically. The key lies in how we apply AI: as a tool to support and complement human expertise, not as a replacement for it. By establishing clear guidelines, maintaining transparency, and ensuring human oversight, we can use AI as a force for good—one that supports the growth and development of knowledge while upholding the ethical standards that define scholarly publishing.

Call to Action

The scholarly publishing community must work together to establish ethical frameworks and guidelines for AI usage. Let’s ensure that AI enhances, rather than undermines, the integrity of research. By being transparent about how AI is applied and retaining human oversight, we can foster an environment where innovation and integrity coexist, ultimately benefiting the advancement of knowledge.

About the author:

Abdul Hakkim is the Senior Manager, Peer Review Services at Integra, a leader in scholarly publishing services. With extensive experience in setting up teams and ensuring the highest standards of quality, Hakkim excels at meeting publisher requirements and driving operational excellence. His expertise spans across supporting research integrity, enhancing manuscript screening, and optimizing peer review processes. Hakkim is dedicated to improving efficiency and fostering innovation in the publishing industry, shaping its future through leadership and strategic growth.