The Science of Detecting LLM-Generated Texts

Abstract

The emergence of large language models (LLMs) has resulted in the production of LLM-generated texts that is highly sophisticated and almost indistinguishable from texts written by humans. However, this has also sparked concerns about the potential misuse of such texts, such as spreading misinformation and causing disruptions in the education system. Although many detection approaches have been proposed, a comprehensive understanding of the achievements and challenges is still lacking. This survey aims to provide an overview of existing LLM-generated text detection techniques and enhance the control and regulation of language generation models. Furthermore, we emphasize crucial considerations for future research, including the development of comprehensive evaluation metrics and the threat posed by open-source LLMs, to drive progress in the area of LLMgenerated text detection.

Have AI-Generated Texts from LLM Infiltrated the Realm of Scientific Writing? A Large-Scale Analysis of Preprint Platforms

“…Some common choices are based on the perplexity of the given text (Dhaini et al, 2023; Ghosal et al, 2023; Tang et al, 2023). The general idea behind this approach is that texts generated by LLMs tend to have lower perplexities.…”

“…Detectors that use hidden statistical patterns have become advantageous in this context, as they require no knowledge of the specific LLMs used and little to no training at all. Some common choices are based on the perplexity of the given text (Dhaini et al, 2023;Ghosal et al, 2023;Tang et al, 2023). The general idea behind this approach is that texts generated by LLMs tend to have lower perplexities.…”

Large Language Models: A Comprehensive Survey of its Applications, Challenges, Limitations, and Future Prospects

“…The susceptibility to adversarial attacks [242], ethical concerns [75], difficulty with context-dependent language [249], absence of emotion and sentiment analysis [513], limited multilingual capabilities [514], limited memory [515], lack of creativity [413], and restricted real-time capabilities [489] are also critical concerns. The high costs of training and maintenance, limited scalability, lack of causality, inadequate ability to handle multimodal inputs, limited attention span, limited transfer learning capabilities, insufficient understanding of the world beyond text, inadequate comprehension of human behavior and psychology, limited ability to generate long-form text, restricted collaboration capabilities, limited ability to handle ambiguity, inadequate understanding of cultural differences, limited ability to learn incrementally, limited ability to handle structured data, and limited ability to handle noise or errors in input data [516], [517], [518], [519], [520], [521], [92] are some of the key challenges in safe, responsible, and efficient deployment of LLMs.…”

A survey on deep reinforcement learning for audio-based applications

“…On a similar vein, Tornede et al [61] explore LLMs in the context of Automated Machine Learning (AutoML) techniques, discussing existing methodologies and the challenges of using them to enhance LLM performance. Tang et al [62] focus on techniques for detecting text generated by LLMs, while Chang et al [29] have examined the various ways to evaluate LLMs. Additionally, there are a number of surveys dedicated to investigating the specialised applications of Large Models in various fields such as vision [23,24,32,33], education [34][35][36][37]63], healthcare [38,39], computational biology [42,43], computer programming [64,65], law [44][45][46]66], or robotics [47,67,68] among others.…”