Detecting Machine-Generated Texts: Not Just "AI vs Humans" and Explainability is Complicated

TL;DR

This paper introduces a ternary classification scheme for detecting machine-generated texts, emphasizing explainability and the importance of an 'undecided' category to improve understanding and trust in detection results.

Contribution

It proposes a novel ternary classification approach for LLM-generated text detection, including datasets, analysis of detector explainability, and guidelines for more transparent detection systems.

Findings

The 'undecided' category enhances explainability in detection.

Top detectors struggle with the 'undecided' class, highlighting explainability challenges.

New datasets reveal the complexity of distinguishing machine from human texts.

Abstract

As LLMs rapidly advance, increasing concerns arise regarding risks about actual authorship of texts we see online and in real world. The task of distinguishing LLM-authored texts is complicated by the nuanced and overlapping behaviors of both machines and humans. In this paper, we challenge the current practice of considering LLM-generated text detection a binary classification task of differentiating human from AI. Instead, we introduce a novel ternary text classification scheme, adding an "undecided" category for texts that could be attributed to either source, and we show that this new category is crucial to understand how to make the detection result more explainable to lay users. This research shifts the paradigm from merely classifying to explaining machine-generated texts, emphasizing need for detectors to provide clear and understandable explanations to users. Our study involves creating four new datasets comprised of texts from various LLMs and human authors. Based on new datasets, we performed binary classification tests to ascertain the most effective SOTA detection methods and identified SOTA LLMs capable of producing harder-to-detect texts. We constructed a new dataset of texts generated by two top-performing LLMs and human authors, and asked three human annotators to produce ternary labels with explanation notes. This dataset was used to investigate how three top-performing SOTA detectors behave in new ternary classification context. Our results highlight why "undecided" category is much needed from the viewpoint of explainability. Additionally, we conducted an analysis of explainability of the three best-performing detectors and the explanation notes of the human annotators, revealing insights about the complexity of explainable detection of machine-generated texts. Finally, we propose guidelines for developing future detection systems with improved explanatory power.