The numerous problems of traditional peer review are familiar to virtually all scientists. However, the answer to critical questions has always been: Is there anything better? Today, the answer could be: "qed Science" (from quod erat demonstrandum , which was to be demonstrated), an AI analysis tool for manuscripts developed by Oded Rechavi and his team.

As early as 2020, Rechavi said in an interview in Current Biology about peer review: "Perhaps a future AI will eventually be advanced enough to be able to help." This point seemed to have been reached shortly after the publication of ChatGPT and other large-language models. In any case, Oded Rechavi and the physician Niv Samuel Mastboim from the Technion in Haifa, along with a small team, set about developing an AI tool that would analyze manuscripts automatically, constructively, fairly, and quickly.

The first result of this project is the publicly accessible AI agent qed Science , which logically structures a submitted manuscript or paper, whether published or unpublished, into claims, tests the proposed structure for validity and potential gaps, and suggests ways to close these gaps. Authors can reject or comment on the gaps ("Doesn't seem to be a gap"), allowing qed Science to continuously learn and improve its analyses. A gap itself can be categorized as "minor" or "major" and is always introduced with a few sentences of background information. Finally, qed Science describes the entire problem with the manuscript and offers two ways to address it: firstly, with specific text changes, and secondly, with suggestions for further experiments or alternative or additional analyses of the data.