Skip to content
NewA calmer, faster ZeroHour

Author-style Gemini rewrites drop QuillBot AI-text scores to a 0% median

Original title: Who Can You Hide Behind? Author-Conditioned Style Transfer and the Evasion of a Commercial AI-Text Detector

arXiv cs.AI / cs.LG / cs.CLby Mohd Ruhul Ameen ()1 min read
AI summary · x-ai/grok-4.7Summaries by Vani AI

Of 991 rewrites prompted with samples from 55 authors, 66% scored exactly 0% on QuillBot.

Researchers rewrote 991 AI-generated answers with Gemini using a writing sample from one of 55 named authors, then scored originals and rewrites with the QuillBot AI detector. The median score fell from 100% to 0% and the mean from 67.9% to 11.2%. Among 699 originals scored at least 50%, 71% of rewrites scored below 10%. Scores varied sharply by the author used in the prompt.

Unconfirmed: Results cover only QuillBot and Gemini and lack a generic-rewrite baseline.

Key facts
  • All rewrites associated with 10 authors scored 0%.
  • Mean score stayed above 50% for six authors.
  • Rewrites with under 5% five-word reuse still averaged 15.2%.
  • Meaning preservation was not measured.
  • Rewrites were substantially longer than the originals.
Explain this AI news with Vani AIWhat was announced, what is new and why it matters, in a new chat

From the original article

1 min read at arxiv.org

AI-text detectors can be sensitive to changes in writing style. We examined this sensitivity by rewriting 991 AI-generated answers with Gemini, using prompts that included a writing sample from one of 55 named authors. We scored each original and its rewrite with the QuillBot AI detector.

arXiv cs.AI / cs.LG / cs.CL

By Mohd Ruhul Ameen · arxiv.org

Read the full article at arXiv cs.AI / cs.LG / cs.CL

A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.