How Close is ChatGPT to Human Experts? Comparison Corpus, Evaluation, and Detection

TL;DR

This paper introduces HC3, a large dataset comparing ChatGPT and human responses across various domains, analyzes their differences, and develops detection methods to distinguish AI-generated from human text.

Contribution

It provides the HC3 dataset for evaluating ChatGPT versus human responses and proposes effective detection systems for AI-generated text.

Findings

ChatGPT responses differ significantly from human experts in style and content.

Detection systems can effectively identify AI-generated text with high accuracy.

Linguistic and content analysis reveal key gaps between ChatGPT and human responses.

Abstract

The introduction of ChatGPT has garnered widespread attention in both academic and industrial communities. ChatGPT is able to respond effectively to a wide range of human questions, providing fluent and comprehensive answers that significantly surpass previous public chatbots in terms of security and usefulness. On one hand, people are curious about how ChatGPT is able to achieve such strength and how far it is from human experts. On the other hand, people are starting to worry about the potential negative impacts that large language models (LLMs) like ChatGPT could have on society, such as fake news, plagiarism, and social security issues. In this work, we collected tens of thousands of comparison responses from both human experts and ChatGPT, with questions ranging from open-domain, financial, medical, legal, and psychological areas. We call the collected dataset the Human ChatGPT Comparison Corpus (HC3). Based on the HC3 dataset, we study the characteristics of ChatGPT's responses, the differences and gaps from human experts, and future directions for LLMs. We conducted comprehensive human evaluations and linguistic analyses of ChatGPT-generated content compared with that of humans, where many interesting results are revealed. After that, we conduct extensive experiments on how to effectively detect whether a certain text is generated by ChatGPT or humans. We build three different detection systems, explore several key factors that influence their effectiveness, and evaluate them in different scenarios. The dataset, code, and models are all publicly available at https://github.com/Hello-SimpleAI/chatgpt-comparison-detection.