QED Science AI tool assesses scientific preprint quality
TL;DR. QED Science developed an AI tool that evaluates life-sciences preprints for originality and validity, claiming to identify the top 1% of research papers. - The AI assesses claims based on presented data and identifies gaps, using negative results to improve its evaluation. - The tool has been piloted by openRxiv on bioRxiv, sparking debate among researchers regarding AI's role in judging scientific quality. - Over 10,000 laboratories have used the free tool, highlighting its growing adoption in academic circles.
- QED Science developed an AI tool to assess the originality and validity of life-sciences preprints.
- The AI platform is trained on diverse data sources, including open reviews and synthetic data, to understand supportive and non-supportive evidence.
- The tool identifies gaps in research and considers negative results to reduce bias towards positive findings.
- openRxiv is piloting the system on bioRxiv, and QED Science used it to rank over 57,000 preprints, identifying a 'top 1%'.
- The ranking has generated debate among researchers concerning AI's reliability in judging scientific quality and potential for academic prestige reinforcement.