Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges

Area of Science:

• Digital Forensics
• Artificial Intelligence
• Natural Language Processing

Background:

• Accurate authorship attribution is vital for digital content integrity, forensic investigations, and combating misinformation and plagiarism.
• The rise of Large Language Models (LLMs) complicates traditional authorship attribution, blurring lines between human and machine-generated text.
• Establishing credibility and accountability in authorship is essential in the digital age.

Purpose of the Study:

• To provide a comprehensive literature review on authorship attribution in the context of Large Language Models (LLMs).
• To systematically explore and categorize current research challenges in authorship attribution.
• To identify future research directions and open problems in this rapidly evolving field.

Main Methods:

• Systematic literature review of recent research on authorship attribution.
• Categorization of attribution problems into four key areas: Human-written Text Attribution, LLM-generated Text Detection, LLM-generated Text Attribution, and Human-LLM Co-authored Text Attribution.
• Discussion of challenges related to generalization and explainability of attribution methods.

Main Results:

• The study categorizes authorship attribution into four distinct problem types.
• Key challenges in generalization across domains and explainability of attribution models are highlighted.
• Existing methods and benchmarks are evaluated to identify limitations.

Conclusions:

• The literature review serves as a roadmap for researchers and practitioners in authorship attribution.
• Further research is needed to address generalization and explainability challenges for robust attribution.
• The study identifies critical open problems and future research directions in the era of LLMs.