Distinguishing ChatGPT(-3.5, -4)-generated and human-written papers through Japanese stylometric analysis

TL;DR

This study analyzes Japanese stylometric features to distinguish texts generated by GPT-3.5 and GPT-4 from human-written papers, demonstrating high classification accuracy using machine learning techniques.

Contribution

It introduces a novel Japanese stylometric analysis approach and demonstrates effective classification of AI-generated versus human texts with high accuracy.

Findings

Distinct stylometric distributions for GPT and human texts identified.

Random forest classifier achieved 100% accuracy in distinguishing AI from human texts.

GPT-4's increased parameters do not make its texts more human-like in stylometric features.

Abstract

In the first half of 2023, text-generative artificial intelligence (AI), including ChatGPT, equipped with GPT-3.5 and GPT-4, from OpenAI, has attracted considerable attention worldwide. In this study, first, we compared Japanese stylometric features of texts generated by GPT (-3.5 and -4) and those written by humans. In this work, we performed multi-dimensional scaling (MDS) to confirm the distributions of 216 texts of three classes (72 academic papers written by 36 single authors, 72 texts generated by GPT-3.5, and 72 texts generated by GPT-4 on the basis of the titles of the aforementioned papers) focusing on the following stylometric features: (1) bigrams of parts-of-speech, (2) bigram of postpositional particle words, (3) positioning of commas, and (4) rate of function words. MDS revealed distinct distributions at each stylometric feature of GPT (-3.5 and -4) and human. Although GPT-4 is more powerful than GPT-3.5 because it has more parameters, both GPT (-3.5 and -4) distributions are likely to overlap. These results indicate that although the number of parameters may increase in the future, GPT-generated texts may not be close to that written by humans in terms of stylometric features. Second, we verified the classification performance of random forest (RF) for two classes (GPT and human) focusing on Japanese stylometric features. This study revealed the high performance of RF in each stylometric feature: The RF classifier focusing on the rate of function words achieved 98.1% accuracy. Furthermore the RF classifier focusing on all stylometric features reached 100% in terms of all performance indexes (accuracy, recall, precision, and F1 score). This study concluded that at this stage we human discriminate ChatGPT from human limited to Japanese language.