The traditional foundation of scientific integrity is showing signs of structural failure. For decades, the peer-review process has acted as the gatekeeper of human knowledge, ensuring that research is vetted by experts before it reaches the public. However, the system is currently buckling under the weight of an unprecedented surge in manuscript submissions, many of which are now being augmented or entirely generated by artificial intelligence. As the volume of papers reaches record highs, the pool of volunteer reviewers is shrinking, leading to a bottleneck that threatens to stall global scientific progress.
What happened
The academic community is reporting a growing crisis where the demand for expert validation far outstrips the supply of available reviewers. Researchers are finding themselves trapped in a cycle of “review or perish,” yet many are opting out of the volunteer-based system due to professional burnout. The situation has been significantly exacerbated by the emergence of large language models (LLMs), which allow authors to produce and submit papers at a velocity never seen before.
This influx has led to a “clogged” system where even high-quality research can sit in limbo for years. In some cases, editors are struggling to find even a single qualified individual willing to critique a paper. The result is a surge in desk rejections and a rise in “paper mills”—unethical organizations that churn out fraudulent research for a fee—which further drain the time and patience of the remaining legitimate reviewers.
Context
Peer review has historically relied on a “gift economy.” Scientists review the work of their colleagues for free, under the assumption that others will do the same for them. However, this model was designed for an era when scientific output was a fraction of what it is today. In the modern “publish or perish” academic landscape, career advancement is tied strictly to the quantity of published work, incentivizing researchers to prioritize writing over reviewing.
The integration of AI into this ecosystem has created a double-edged sword. While AI tools can help researchers summarize data or polish their prose, they also make it easier to flood journals with low-effort or fabricated studies. Compounding the issue is the fact that reviewers themselves are increasingly using AI to write their evaluations. This creates a “dead internet” scenario for academia, where AI-generated papers are being critiqued by AI-generated reviews, potentially bypassing the human oversight that is critical for verifying complex scientific claims.
Why it matters
The collapse of peer review has implications far beyond the walls of a laboratory. If the system fails to filter out flawed or fraudulent research, the public’s trust in science could be permanently damaged. We have already seen how misinformation regarding public health and climate change can have catastrophic real-world consequences; a weakened peer-review system only makes it easier for such errors to be codified as “official” science.
Furthermore, the delay in publishing critical research slows down innovation. When life-saving medical discoveries or technological breakthroughs are stuck in a multi-year backlog, society as a whole loses. Moving forward, the academic community must decide if the volunteer-based model is still viable or if fundamental changes—such as paying reviewers or developing more sophisticated AI-detection protocols—are necessary to save the integrity of human knowledge. Without a sustainable solution, the very mechanism that ensures scientific truth may become a relic of the past.