Segmenting Human-LLM Co-authored Text via Change Point Detection

2026-05-05

AI Summary

This study addresses the limitation of existing human/LLM text detection methods, which are restricted to whole-segment binary classification and cannot localize individual contributions in human-AI collaborative writing. The work reframes this challenge as a change-point detection problem in time series, leveraging statistical properties of LLM-generated text. It introduces weighted and generalized change-point detection algorithms that effectively handle heterogeneity in detection scores and establishes their minimax optimality. Experimental results demonstrate that the proposed approach significantly outperforms multiple baselines in segmentation accuracy, enabling high-precision localization of human and machine-authored segments within co-authored texts.

Abstract

The rise of large language models (LLMs) has created an urgent need to distinguish between human-written and LLM-generated text to ensure authenticity and societal trust. Existing detectors typically provide a binary classification for an entire passage; however, this is insufficient for human--LLM co-authored text, where the objective is to localize specific segments authored by humans or LLMs. To bridge this gap, we propose algorithms to segment text into human- and LLM-authored pieces. Our key observation is that such a segmentation task is conceptually similar to classical change point detection in time-series analysis. Leveraging this analogy, we adapt change point detection to LLM-generated text detection, develop a weighted algorithm and a generalized algorithm to accommodate heterogeneous detection score variability, and establish the minimax optimality of our procedure. Empirically, we demonstrate the strong performance of our approach against a wide range of existing baselines.

Problem

• human-LLM co-authored text
• text segmentation
• change point detection
• LLM-generated text
• authenticity

Innovation

• change point detection
• human-LLM co-authored text
• text segmentation
• minimax optimality
• heterogeneous detection scores