Who should still peer-review, for whom, and with what tools?

Stefano Mammola , Lorenzo Vilizzi , Paride Balzani, Danish Ali Ahmed , Lucian Pârvulescu, Joana Garrido Nogueira, Manuel Lopes-Lima, Ronaldo Sousa, Mattias L Johansson, Alessandro Nota, Carlos Cano-Barbacil, Ali Serhan Tarkan, Eva Záhorská, Anna Gavioli, Virginie Raybaud, Emili Garcia-Berthou, Baran Yoğurtçuoğlu, Jingrui Sun , Ross N. Cuthbert, Elizabeta Briski, Ester M Eckert, Barbora Števove, Teun Everts, Diego Fontaneto, Stelios Katsanevakis, Riccardo Scalera, Michel Bariche, Valentini Stamatiadou, Marta Rodríguez-Rey, Phillip Joschka Haubrock 

In drafting On the Origin of Species, Charles Darwin developed his thoughts through extensive correspondence with trusted colleagues, seeking feedback on ideas, manuscripts, and proofs. This informal peer-review process, like many others of its time, was based on personal trust, academic respect, and reputation rather than guided by any institution, and it helped pave the way for modern peer review. Today, peer review has been largely corrupted by corporate greed: a handful of dominant mega-publishers exploit reviewers (and to an extent also editors) as an unpaid, external workforce expected to supply scientific judgement, quality control, and career signalling that journals package and sell but do not themselves produce. Amid an ever-increasing volume of review requests reaching researchers' inboxes daily, many respond by declining invitations outright, if they respond at all. Others accept but rely heavily on AI, since this saves time and a few more reviews will look good on their Web of Science profile. Thorough intellectual assessments are becoming increasingly rare. The current system is not sustainable, and change is urgently needed before expert judgement is fully displaced by speed, prestige signalling, and machine-written reports. Solutions include distributing unpaid peer-review labour more wisely, prioritizing non-commercial publishers and academia-friendly journals, exploring alternative business models, and redesigning peer review so that human expertise is concentrated on tasks requiring scientific judgement while low-order tasks are supported by AI. Above all, researchers, editors, funders, publishers, and societies should stop treating review as free certification labour and instead recognize it as a scarce scholarly resource that must be used where it contributes most to scientific quality.