Forestry journals are lacking peer reviewers and could AI fix this?
image: Radar chart comparing emphasis placed on six review categories by a human Associate Editor with two independent peer reviewers and AI large language models. Both human reviewers and all four AI models identified poor model fit and inconsistent AUC values as primary weaknesses, suggesting AI can reliably detect certain methodological issues
Credit: Matthew B. Russell and Aaron R. Weiskittel
A new study finds that forestry research articles have quadrupled since 2000, while the pool of available reviewers has shrunk, creating a growing crisis in how science is evaluated before publication.
Published in Forest Ecosystems, the commentary by Matthew Russell and Aaron Weiskittel draws on data from 156 journals over 25 years. Articles published annually have risen from around 3,500 in 2000 to over 14,000 in 2024, while the US forestry research workforce has shrunk by 12% and reviewer decline or non-response rates have been recorded nearly as high as 48%. The authors project the demand-to-supply ratio for reviewers could reach 9.4:1 by 2032.
Open-access publishing has compounded the pressure. Now accounting for 45% of forestry articles (up from 20% in 2000), open-access journals often ask reviewers to turn around assessments in as few as 10 days rather than the standard three weeks.
As AI has become an everyday tool for us, the question is: can AI be of assistance in peer review, and how well does it perform? Comparing reviews of the same forest biometrics-focused paper by human experts and four AI models (ChatGPT, Claude, Gemini, and Perplexity), they found that AI reliably catches statistical inconsistencies and presentation issues, but falls short of the domain-specific insight that human reviewers provide, particularly in evaluating how well models perform across species, regions, and tree sizes. In peer review, AI models highlighted presentation of results while human reviewers emphasized the importance of model verification. With 53% of researchers now reporting AI use in peer review, the authors argue the question is no longer whether AI will be integrated, but how responsibly it will be done.
"The goal should not be to replace peer review but to evolve it," the authors write. They recommend that journals should pilot pre-publication review, introduce version control for manuscripts, offer modest compensation for reviewers, and develop clear journal policies that distinguish appropriate AI tasks from those requiring human judgment.
Rather than viewing the current pressure as purely a crisis, Russell and Weiskittel argue that the proliferation of forestry journals creates a rare opportunity. "By treating peer review as a system amenable to evidence-based improvement rather than a fixed tradition, forestry journals can maintain scientific rigor while adapting to contemporary realities."
DOI Link: