A Cross-Disciplinary Academic Evaluation of Generative AI Models in HR, Accounting, and Economics: ChatGPT-5 vs. DeepSeek

Najib Bou Zakhem (), Malak Bou Diab and Suha Tahan
Additional contact information
Najib Bou Zakhem: Management & International Management Department, School of Business, Lebanese International University, Bekaa 146404, Lebanon
Malak Bou Diab: Accounting Information Systems Department, School of Business, Lebanese International University, Beirut 146404, Lebanon
Suha Tahan: Economics Department, School of Business, Lebanese International University, Bekaa 146404, Lebanon

Administrative Sciences, 2025, vol. 15, issue 11, 1-22

Abstract: As generative AI is being further integrated into academic and professional contexts, there is a demonstrable need to determine the performance of generative AI within specific, applied domains. This research compares the performances of ChatGPT-5 and DeepSeek on tasks in the domains of accounting, economics, and human resources. The models were provided two prompts per domain, and outputs were evaluated by academics across five criteria: accuracy, clarity, conciseness, systematic reasoning, and indicators of potential bias. The inter-rater reliability was reported using Cohen’s Kappa. From the findings, both models display differences in performance. ChatGPT-5 outperformed DeepSeek in accounting and human resources, while DeepSeek outperformed ChatGPT-5 on epistemic economics tasks. Since results have shown that ChatGPT-5 outperformed DeepSeek in two out of three domains, the research recommends a reliability-based framework to compare generative AI outputs within business disciplines and offers practical suggestions on when and how to use the models within academic and professional contexts.

Keywords: generative AI; ChatGPT; DeepSeek; education; technology; human resources; accounting; economics; artificial intelligence (search for similar items in EconPapers)
JEL-codes: L M M0 M1 M10 M11 M12 M14 M15 M16 (search for similar items in EconPapers)
Date: 2025
References: Add references at CitEc
Citations:

Downloads: (external link)
https://www.mdpi.com/2076-3387/15/11/412/pdf (application/pdf)
https://www.mdpi.com/2076-3387/15/11/412/ (text/html)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:gam:jadmsc:v:15:y:2025:i:11:p:412-:d:1778653

Access Statistics for this article

Administrative Sciences is currently edited by Ms. Nancy Ma

More articles in Administrative Sciences from MDPI
Bibliographic data for series maintained by MDPI Indexing Manager ().