<p>Artificial intelligence (AI) is transforming scientific research, but its application in peer review raises ethical and methodological concerns. This paper examines the risks of AI-driven peer review, focusing on the lack of domain expertise among reviewers, AI’s ability to challenge any article when stress-tested, and the potential compromise of research integrity. These issues compound systemic problems such as self-citations, citation markets, lenient reviews, and the proliferation of special issues that grant guest editors, especially in high-impact journals, undue influence over academic careers. Predatory journals publishing thousands of papers with questionable review processes further threaten the credibility of scientific databases, creating conflicts of interest and financial incentives within academic publishing. An overlooked factor worsening these challenges is that peer reviewers are unpaid, making the process unsustainable. As scholars struggle to balance reviewing with their own research, many rely on AI tools to expedite evaluations, potentially sacrificing depth and rigor. While AI can assist, its overuse risks diminishing the reliability of scholarly publishing. If high-quality peer review is to be preserved, the academic community must reconsider incentives and structural reforms to ensure fair and thorough assessments. This paper illustrates how AI can fabricate pretextual negative reviews, enabling the rejection of even high-quality, previously published papers. Through a concrete example, we show how AI-driven review systems, when manipulated, can selectively generate biased assessments to serve editorial or institutional agendas. Addressing these vulnerabilities requires a fundamental reevaluation of peer review practices to maintain research integrity and safeguard the future of scientific publishing.</p>