Peer review, the backbone of scientific validation, is grappling with an outdated and inefficient reviewer selection process. Editors face mounting challenges in identifying suitable experts, often relying on limited tools or author-supplied references that can introduce bias. Professor Philipp Koellinger, co-founder and CEO of DeSci Labs, argues that artificial intelligence (AI) can revolutionise this critical step, enhancing fairness, speed, and transparency in scientific publishing.
The struggle to find the right reviewers
Selecting peer reviewers is a cornerstone of ensuring research quality, yet the process remains slow and flawed. Editors typically scan reference lists or rely on databases of past referees, often reusing the same reviewers. A 2023 survey across four scientific journals revealed that editors found identifying multiple suitable reviewers the most frustrating aspect of their role, with many struggling to pinpoint experts best suited for each manuscript.
This reliance on a limited pool risks inadequate evaluations. Reviewers may lack the precise expertise or impartiality needed, leading to biased or subpar assessments. Despite technological advances, many journals cling to these aging systems, undermining the diversity and reliability of peer review, which directly impacts the integrity of published science and funding decisions.
Qualities of an ideal reviewer
An effective peer reviewer must meet strict criteria to ensure rigorous and fair evaluations:
- No conflicts of interest. Reviewers should not be recent co-authors, supervisors, supervisees, or affiliated with the same institution as the authors.
- Relevant expertise. They must have recent publications on the topic or use similar methods, ensuring a deep understanding of the research.
- Objectivity. Reviewers must set aside personal biases or rivalries to provide impartial judgments.
- Thoroughness. They should be motivated to dedicate time to carefully assess submissions.
These qualities are essential for maintaining trust in scientific research. Inappropriate reviewers can introduce bias, reduce review quality, and erode the credibility of the publishing process.
Flaws in current methods
Editors and funding agencies spend significant time – often up to two weeks – identifying and contacting reviewers within a specific field. Many rely heavily on a manuscript’s reference list to save time, but this approach has drawbacks. Authors can manipulate citations to steer suggestions toward preferred reviewers, while uncited experts may be overlooked, narrowing the reviewer pool.
This method often results in biased or incomplete evaluations, missing broader expertise that could elevate review quality. Additionally, peer reviewers typically work voluntarily, making it harder to secure timely and thorough assessments. These challenges delay publications and compromise the rigour of scientific evaluation, as editors juggle demanding workloads with limited resources.
AI as a game-changer
AI-powered tools offer a transformative solution to these long-standing issues. By rapidly analysing vast amounts of scientific literature, AI can identify the most relevant experts for each manuscript. Advanced systems, like DeSci Labs’ Referee Finder, incorporate features such as:
- Conflict-of-interest checks. Ensuring reviewers have no ties to the authors.
- Expertise mapping. Distinguishing between topic specialists and methodological experts.
- Researcher metadata. Including H-index and career stage for a comprehensive match.
These tools expand the reviewer pool beyond traditional reference-based methods, reducing bias and uncovering overlooked experts. Editors can customise algorithms to align with each submission’s unique needs, ensuring precise matches. This approach not only speeds up the process but also enhances fairness and transparency in peer review.
A path to fairer science
AI-driven solutions address inefficiencies that have plagued peer review for decades. By moving beyond manipulable references and limited databases, these tools open the door to a more robust and equitable publishing process. As Professor Koellinger notes: “AI can help science live up to its ideals,” ensuring that the best experts evaluate research with impartiality and precision.
The adoption of AI in reviewer selection could mark a turning point for scientific publishing. With tools like Referee Finder, journals can uphold the integrity of peer review while meeting the demands of a rapidly evolving research landscape. As science grows more complex, embracing AI is not just an upgrade—it’s a necessity for maintaining trust in the global scientific enterprise.
