Every other day, I hear the same concern from colleagues and friends: peer-review reports have become increasingly generic since the rise of AI tools such as ChatGPT.
Many researchers feel that reviewer comments today are less intellectual, less insightful, and more like standardized checklists than genuine scientific critiques. Some authors have even found strong indications that reviewer reports were AI-generated, despite many journals now explicitly requiring reviewers not to use generative AI to evaluate confidential manuscripts.
The obvious question is: Why is this happening?
Many argue that AI is destroying the traditional peer-review system. But I believe AI is merely exposing problems that have existed for decades.
One legitimate concern is confidentiality. Authors often spend years developing ideas, collecting data, and writing manuscripts before submission. If a reviewer uploads an unpublished manuscript into an external AI system, there is a potential risk of violating confidentiality and research integrity. This concern is real and deserves serious attention.
However, we should also ask ourselves an uncomfortable question:
Was peer review truly functioning perfectly before AI?
The answer is probably no.
Traditional peer review has always relied on voluntary contributions from busy scientists. Long before generative AI existed, many principal investigators (PIs), overwhelmed by their responsibilities, routinely delegated manuscript reviews to junior faculty members, postdoctoral researchers, or PhD students. In many cases, these junior researchers performed the actual review without any formal acknowledgment from the journal. While this often-provided valuable training, it also raised important questions about transparency and research integrity.
Another long-standing problem was conflict of interest. There have been cases where manuscripts were reviewed within competing laboratories, allowing reviewers to delay publication while advancing similar work in their own groups. AI did not create these ethical problems—they already existed.
Today, the workload on researchers is greater than ever. A modern PI is expected to:
- Write multiple grant proposals.
- Manage a research group.
- Teach and mentor students.
- Handle increasing administrative responsibilities.
- Publish continuously to satisfy institutional metrics and global rankings.
- And, on top of all this, provide thoughtful, detailed peer reviews—entirely voluntarily.
Under such pressure, it is perhaps unsurprising that some reviewers turn to AI tools to save time. This does not justify the practice, particularly when confidentiality may be compromised. But simply blaming AI ignores the deeper structural issues.
The real question is not "Who is using AI?" but "Why has the system reached a point where reviewers feel they need AI to complete their reviews?"
At the same time, authors are increasingly paying substantial Article Processing Charges (APCs) to publish their work, while publishers continue to rely on unpaid reviewers whose expertise is essential for maintaining scientific quality.
Can this model remain sustainable?
If publishers generate significant revenue from the publication process, should some of that value be reinvested into the peer-review system? Should reviewers receive recognition, protected time, institutional credit, or even financial compensation for high-quality reviews?
These are difficult questions, but they can no longer be ignored.
Peer review remains one of the foundations of scientific integrity. However, integrity cannot depend indefinitely on goodwill alone while academic workloads continue to rise.
Blaming AI is easy.
Fixing the incentives that have weakened peer review for years is much harder—but that is where the real solution lies.
If we fail to reform the system, we risk arriving at a future where thoughtful human scientific critique is gradually replaced by automated, generic reviews. The greatest loss will not simply be the rise of AI—it will be the loss of intellectual engagement that has always been the true value of peer review.
Competing Interests:
The author declares no competing interests. (Poster image is generated by ChatGPT open AI). This content was prepared with editorial assistance from ChatGPT to enhance clarity and presentation. The author takes full responsibility for the content.