TY - GEN
T1 - Semantic Similarity Detection of AI-Generated Academic Content
T2 - 2nd International Generative AI and Computational Language Modelling Conference, GACLM 2025
AU - Odeh, Ayman
AU - Salameh, Haythem Bany
AU - Elrefae, Ghaleb
N1 - Publisher Copyright:
© 2025 IEEE.
PY - 2025
Y1 - 2025
N2 - 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.
AB - 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.
KW - Academic Integrity
KW - Artificial Intelligence
KW - Semantic Similarity
KW - Similarity Metrics
UR - https://www.scopus.com/pages/publications/105026292012
U2 - 10.1109/GACLM67198.2025.11232384
DO - 10.1109/GACLM67198.2025.11232384
M3 - Conference contribution
AN - SCOPUS:105026292012
T3 - 2025 2nd International Generative AI and Computational Language Modelling Conference, GACLM 2025
SP - 198
EP - 202
BT - 2025 2nd International Generative AI and Computational Language Modelling Conference, GACLM 2025
A2 - Lloret, Jaime
A2 - Jararweh, Yaser
PB - Institute of Electrical and Electronics Engineers Inc.
Y2 - 18 August 2025 through 21 August 2025
ER -