Lexical Density and Readability in AI-Generated Academic Writing: A Cross-Disciplinary Corpus Comparison with Human-Authored Texts
Keywords:
lexical density, readability, AI-generated text, large language models, academic writing, corpus analysis, natural language processingAbstract
This study examines differences in lexical density and readability between AI-generated academic texts and human-authored scholarly writing across the disciplines of linguistics, economics, and computer science. As large language models (LLMs) become increasingly integrated into academic writing practices, understanding their linguistic characteristics has become essential for evaluating their impact on scholarly communication and educational policy. Drawing on corpus linguistics and computational text analysis, the study compares a balanced corpus of 300 academic texts comprising 150 AI-generated texts produced using GPT-4 and Claude 3 and 150 human-authored articles drawn from peer-reviewed journals. The corpus contains approximately 600,000 words distributed evenly across the three disciplinary domains. Quantitative analyses were conducted using AntConc and Python-based natural language processing tools to measure lexical density, type-token ratio, Flesch-Kincaid readability scores, and nominalisation frequency. The results indicate significant differences between AI-generated and human-authored writing. AI-generated texts exhibit substantially higher lexical density and greater nominalisation frequency, reflecting a tendency toward compressed information packaging and formal academic style. However, these texts also display lower type-token ratios, suggesting reduced lexical diversity and greater reliance on recurrent vocabulary patterns. Despite their higher lexical density, AI-generated texts achieved significantly higher readability scores, indicating that they are generally easier to read than human-authored texts. This apparent paradox is explained by the shorter sentence lengths, more regular syntactic structures, and greater linguistic predictability characteristic of AI-generated prose. The findings highlight distinctive stylistic features of machine-generated academic discourse and raise important questions regarding authorship, academic integrity, writing pedagogy, and the future role of artificial intelligence in scholarly communication. The study contributes to emerging research on AI writing and corpus-based analyses of digital academic discourse.
References
Ahmed, S. (2025). Discursive Constructions in Macbeth: A Systemic-Functional Contribution to English Language Teaching. Academy of Education and Social Sciences Review, 5(4), 568-578.
Al Asbahi, A. A. M. H., Fang, Z., Chandio, Z. A., Tunio, M. K., Ahmed, J., & Abbas, M. (2020). Assessing barriers and solutions for Yemen energy crisis to adopt green and sustainable practices: A fuzzy multi-criteria analysis. Environmental Science and Pollution Research, 27(29), 36765-36781.
Al-Hawtali, A. M., Zaib, K., & Al-Hossini, A. S. (2025). Voices in exile: Postcolonial identity and Muslim immigrant experience in Abdulrazak Gurnah's Admiring Silence. Academy of Education and Social Sciences Review, 5(4), 579-588.
Ali, U. A., Rafique, M., Dahri, A. S., & Kolachi, I. A. (2025). Breaking the vicious cycle: Gender empowerment, political stability, and anti-corruption as catalysts for national statistical reforms. Policy Journal of Social Science Review, 3(10), 234-259.
Aziz, A., Rafique, M., Azher, S., & Ali, M. (2025). Examining the Effectiveness of Women Business & Law Empowerment: A Global Cross Sectional Analysis. Research Consortium Archive, 3(4), 239-277.
Brahmi, M., Hussain, Z., Majeed, M. U., Khan, A., Qureshi, M. A., & Bansal, R. (2025). Corporate social responsibility’s influence on brand image in the automotive sector: the corporate reputation and product quality role. Administrative Sciences, 15(4), 121. https://doi.org/10.3390/admsci15040121
Dahri, A. S., Umrani, A. I., Asif, M., & Mehmood, S. (2025). The Evolution of the Psychological Contract in the Gig Economy. Journal of Management & Social Science, 2(6), 131-149.
Ghani, H., Khan, C., Miraj, J., & Shaikh, J. A. (2023). GREEN MARKETING AND ITS IMPACT ON CONUSMER PURCHASING BEHAVIOR OF BALOCHISTAN. THE ASIAN BULLETIN OF GREEN MANAGEMENT AND CIRCULAR ECONOMY, 3(1), 122-142.
Hussain, A., Ahmad, J., Altaf, Y., Khan, M., & Sanghi, A. (2016). A study on various attributing reasons for switching telecommunication brands among various cellular services at Larkana, Sindh, Pakistan. Finance and Management Sciences, 11(1), 36-83.
Katper, N., Salman, S. M., Mohy‐ud‐Din, K., & Hassan, M. (2025). Mapping Interference in Green Finance, Environmental, and Social Innovation: A New Insight From Quantile‐on‐Quantile Approach. Environmental Quality Management, 35(1), e70163.
Khan, A., Ahmed, J., & Khan, F. U. (2019). An Empirical Study of Customer Satisfaction and Loyalty: An Evidence of Telecom Sector in Pakistan. Journal of Business & Tourism, 5(2), 65-78.
KHAN, A., LIAQUAT, S., SHEIKH, J., & PIRZADO, A. A. (2021). The impact of corona virus (COVID-19) on E-business in Pakistan. Journal of Contemporary Issues in Business and Government| Vol, 27(3), 83.
Khan, R. A., Bibi, M., Khokhar, M., & Tufail, M. M. B. (2025). Geographic Information System and AI Integration to Support Sustainable Environment: A Glance from the Construction Sector. In Harnessing AI in Geospatial Technology for Environmental Monitoring and Management (pp. 145-176). IGI Global Scientific Publishing.
Rafique, M. O. (2025). Impact of the Shariah Board's Characteristics on the Financial Soundness of Islamic Banks. In The Routledge Handbook of Islamic Economics and Finance (pp. 320-329). Routledge.
Rafique, M. O., & Raza, K. M. (2025). Islamic Green Finance: A Marriage of Ethics and Environmental Responsibility in Light of Maqasid al-Shari'ah. In Islamic Green Finance (pp. 3-19). Routledge.
Rafique, M. O., Hureri, M. A., & Riaz, M. (2025). Green banking: Performance of three Pakistani Islamic banks in environmental projects. Southern Journal of Social Sciences, 3(1). Retrieved from https://sjss.isp.edu.pk/index.php/about/article/view/55/31
Shaikh, D. H., Rafique, M., Dahri, A. S., & Kolachi, I. A. (2025). From inequality to integrity: An empirical investigation into the determinants of regulatory quality. Journal of Social Horizons, 2(4), 464-473. https://doi.org/10.5281/zenodo.17335087
Shaikh, D. H., Rafique, M., Dahri, A. S., & Kolachi, I. A. (2025). From inequality to integrity: An empirical investigation into the determinants of regulatory quality. Journal of Social Horizons, 2(4), 464-473.
Shaikh, J. A., & Hussain, A. (2020). An Empirical Study Of Consumer Perceptions Towards Online Shopping.
Shaikh, P. A., Ahmed, M., Yousaf, H., & Ahmed, J. (2020). The determinants of poverty: A case study of district Lasbela, Balochistan, Pakistan. International Journal of Advanced Science and Technology, 29(7), 9688-9700.
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2026 RIDSCON – International Conference on Research Innovations in Social Sciences
This work is licensed under a Creative Commons Attribution 4.0 International License.
