Machine Text Detectors are Membership Inference Attacks

TL;DR

This paper explores the theoretical and empirical connections between membership inference attacks and machine-generated text detection, revealing transferability of methods and introducing a unified evaluation suite.

Contribution

It demonstrates the transferability between MIAs and text detection, unifies methods under an optimal metric, and introduces MINT for cross-task evaluation.

Findings

Strong correlation in cross-task performance (ρ ≈ 0.7)

Machine text detectors perform well on both tasks

Unified evaluation suite MINT implemented with 15 methods

Abstract

Although membership inference attacks (MIAs) and machine-generated text detection target different goals, their methods often exploit similar signals based on a language model's probability distribution, and the two tasks have been studied independently. This can result in conclusions that overlook stronger methods and valuable insights from the other task. In this work, we theoretically and empirically demonstrate the transferability, i.e., how well a method originally developed for one task performs on the other, between MIAs and machine text detection. We prove that the metric achieving asymptotically optimal performance is identical for both tasks. We unify existing methods under this optimal metric and hypothesize that the accuracy with which a method approximates this metric is directly correlated with its transferability. Our large-scale empirical experiments demonstrate very strong rank correlation ($\rho \approx 0.7$) in cross-task performance. Notably, we also find that a machine text detector achieves the strongest performance among evaluated methods on both tasks, demonstrating the practical impact of transferability. To facilitate cross-task development and fair evaluation, we introduce MINT, a unified evaluation suite for MIAs and machine-generated text detection, implementing 15 recent methods from both tasks.