greCAPTCHA prototype aims to check authors’ understanding of their manuscripts

The team says its greCAPTCHA prototype dynamically generates multi-level questions to evaluate an author’s understanding of a submitted manuscript, then produces an automated evaluation report from the answers. It describes testing the prototype through a user study and semi-structured interviews with 31 active researchers. The team reports an AUC score of 0.90 for automated scoring’s ability to distinguish genuine manuscript authors from other participants.