aifollow.news
Back Rohan Paul
Rohan Paul· @rohanpaul_ai · X· · Original publication time SelectedAI score60

University of Maryland study says frontier models can infer authorship from a single textMachine translation

AI introduction

Researchers gave models one text and a zero-shot prompt, without web search, tools or a candidate list, and tested whether they could identify human or AI authorship, the model's lab, or an individual author. The source says Claude Fable 5.2 identified reviewers for 56% of 398 private peer reviews whose authors had never been made public, suggesting that anonymity based on writing alone may be weaker than expected.

Original post · Original

@rohanpaul_ai

– https://arxiv.org/abs/2611.26216

Title: "Frontier models need only one text to unmask its human or AI author"

ReplyRohan Paul@rohanpaul_ai
New University of Maryland paper finds that frontier models now detect AI-written prose as well as trained detectors like Pangram, while also explaining their reasoning. Can name a text's author, AI lab or specific person, from that text alone. Anonymous peer review and pseudonymous accounts assume writing alone doesn't reveal who you are. Spotting AI text was also thought to need trained detectors, not plain prompting. The authors gave models a single text and a zero-shot prompt, with no web search, tools, or candidate list. They tested AI-or-human detection, which lab's model wrote it, and which person wrote it. Claude Fable 5.2 named the reviewer of 56% of 398 private peer reviews whose authors were never made public. It named the account behind 69% of comments from Redditors with over 500 posts in a year. If your product depends on anonymous writing, such as blind review or pseudonymous forums, treat author anonymity as weak and test it against frontier models.
View this post on X ↗

Source: Rohan Paul · x.com

Research
Found an error? Send a correction