The field of AI-generated text detection is rapidly advancing, with a focus on developing innovative methods to identify and distinguish between human-written and machine-generated content. Recent research has explored various approaches, including the use of psycholinguistic features, contrastive learning, and adaptive detection techniques. These advancements have significant implications for maintaining the integrity and authenticity of online content, particularly in domains such as social media, academic publishing, and software development. Noteworthy papers in this area include: RedNote-Vibe, which introduces a longitudinal dataset for social media AI-generated text analysis, and Sci-SpanDet, which presents a structure-aware framework for detecting AI-generated scholarly texts. Additionally, AdaDetectGPT proposes an adaptive detection method with statistical guarantees, and VietBinoculars achieves high accuracy in detecting Vietnamese LLM-generated text.
Detecting AI-Generated Text in Various Domains
Sources
RedNote-Vibe: A Dataset for Capturing Temporal Dynamics of AI-Generated Text in Social Media
Mixture of Detectors: A Compact View of Machine-Generated Text Detection
Active Authentication via Korean Keystrokes Under Varying LLM Assistance and Cognitive Contexts
How Well Do LLMs Imitate Human Writing Style?
CEAID: Benchmark of Multilingual Machine-Generated Text Detection Methods for Central European Languages
Understanding Collective Social Behavior in OSS Communities: A Co-editing Network Analysis of Activity Cascades
VietBinoculars: A Zero-Shot Approach for Detecting Vietnamese LLM-Generated Text
Threats to the sustainability of Community Notes on X
Span-level Detection of AI-generated Scientific Text via Contrastive Learning and Structural Calibration
AdaDetectGPT: Adaptive Detection of LLM-Generated Text with Statistical Guarantees