Dies ist eine Übersichtsseite mit Metadaten zu dieser wissenschaftlichen Arbeit. Der vollständige Artikel ist beim Verlag verfügbar.
Semantic Similarity Detection of AI-Generated Academic Content: A Temporal Analysis Using Machine Learning Techniques
0
Zitationen
3
Autoren
2025
Jahr
Abstract
The rapid evolution of artificial intelligence (AI) technologies has introduced opportunities and challenges in academic research, particularly concerning the generation and authorship of scholarly content. This study investigates the ethical and practical implications of using AI-generated text in academic writing, specifically focusing on research abstracts. A dataset of 18,000 abstracts published between 2014 and 2025 was analyzed. Using only paper titles as prompts, new abstracts were generated via ChatGPT and compared to their original counterparts using multiple similarity metrics, including semantic similarity. The findings indicate a significant rise in similarity scores between years, but most significantly after 2020, reflecting increasing alignment between AI-generated and human-written work. The paper discusses the possible threat posed by the trend to academic integrity and presents guidelines on making transparent, ethical, and accountable uses of AI in academic communication.
Ähnliche Arbeiten
Explainable Artificial Intelligence (XAI): Concepts, taxonomies, opportunities and challenges toward responsible AI
2019 · 8.339 Zit.
Stop explaining black box machine learning models for high stakes decisions and use interpretable models instead
2019 · 8.211 Zit.
High-performance medicine: the convergence of human and artificial intelligence
2018 · 7.614 Zit.
Proceedings of the 19th International Joint Conference on Artificial Intelligence
2005 · 5.776 Zit.
Peeking Inside the Black-Box: A Survey on Explainable Artificial Intelligence (XAI)
2018 · 5.478 Zit.