Source: Latest News from Science Magazine
Part of: AI cracks hard problems in science and math →
A competency inversion is underway: machine learning systems are now more reliable than peer reviewers at catching methodological mistakes in published medical research, where false positives directly translate to harmful clinical decisions. Peer review—the supposed quality gate for medical knowledge—has always depended on human attention and expertise, both of which are scarce and unevenly distributed across journals and specialties. The tension is whether regulatory bodies will accept algorithmic vetting as a substitute for or complement to human scrutiny, and whether this forces a reckoning with systematic underinvestment in reviewing infrastructure.