June 28, 2026
Peer Review or Peer Roast?
Does Your Paper Really Suck?
Scientists are roasting the idea that one AI score can decide if your paper is trash
TLDR: QED Science says its AI can score research papers for quality, but a detailed critique argues the evidence is shaky. In the comments, readers mock the idea of robots grading science and warn it could turn publishing into a pay-to-win score-chasing game.
Science has found a brand-new way to start a fight: let an artificial intelligence tool slap a single number on research papers and hint that anything outside the top 1% basically "sucks." That’s the drama around QED Science’s new "QED score," which uses chatbots to judge whether a paper is original and believable. The company says this number is a faster, better signal than old-school clues like famous journals or big-name universities. Critics reading the review at QED Science are very much not buying the hype.
The comments are where the real sparks fly. One camp agrees the problem is real: there are simply too many papers, too much jargon, and now too much AI-written sludge for anyone to keep up. But as one commenter put it, the diagnosis may be right while the "treatment" is wildly wrong. Another turned the whole thing into a punchline: an AI read a paper, declared it top 1%, and everyone applauded. Ouch.
Then came the capitalism panic. A particularly spicy take warned this could become a subscription trap where researchers feel forced to pay to pre-check and "optimize" their work for a better score, whether the science actually improves or not. Meanwhile, doom-posters called AI the final "1 ton tungsten rod" smashing an already broken peer review system. Translation for non-scientists: people are worried research quality is becoming a popularity contest run by robots, and the comments section is having an absolute field day with it.
Key Points
- •The article reviews QED Science’s QED Score, an LLM-based metric intended to assess scientific paper quality using originality and validity evaluations.
- •QED Science’s white paper claims the QED Score is faster, more accurate, and less biased than journal rank, and supports this claim with three validation studies against SCImago Journal Rank (SJR).
- •The article concludes that the presented evidence does not support claims that the QED Score is a more accurate or less biased measure of scientific quality.
- •In case study 1, QED reportedly outperformed SJR on some AUC-based classification tasks using 975 expert-labeled papers, but the article says the study is methodologically opaque and not reproducible from the published information.
- •In case study 2, the article describes a comparison of QED scores for 2,879 bioRxiv preprints with the SJR of journals where they were later published and characterizes the evidence as inconsistent.