Research
AI Writing on arXiv Measured: One Third of New Papers Flagged
A new study scored the full text of 12,750 arXiv papers and found that roughly one third of recent submissions read as machine-written. The detector was calibrated to a 0.4% false-positive rate on pre-ChatGPT papers. Computer science leads at 65%, while mathematics is lowest at 0.7%, though the authors note limitations in detecting AI writing in notation-heavy fields.
Jul 205 minNeura News