Dies ist eine Übersichtsseite mit Metadaten zu dieser wissenschaftlichen Arbeit. Der vollständige Artikel ist beim Verlag verfügbar.
Evaluating the accuracy of ChatGPT in scientific writing
7
Zitationen
1
Autoren
2023
Jahr
Abstract
Abstract This paper aims to assess the quality of ChatGPT’s answers when used to write a research paper and to explore its potential in assisting researchers in scientific writing. We made several requests to two versions of ChatGPT (3.5 and 4), with each one addressing a specific part of the research paper (Introduction, Literature Review, Data, and Code Implementation). The results indicate that ChatGPT could provide incorrect or even non-existent information for literature review, bibliographic references and data sources. Additionally, the results show that ChatGPT (3.5 and 4) require human assistance and verification of the accuracy of provided information after the fact. Also, this study confirms that ChatGPT-4 is more accurate than ChatGPT-3.5.
Ähnliche Arbeiten
Explainable Artificial Intelligence (XAI): Concepts, taxonomies, opportunities and challenges toward responsible AI
2019 · 8.260 Zit.
Stop explaining black box machine learning models for high stakes decisions and use interpretable models instead
2019 · 8.116 Zit.
High-performance medicine: the convergence of human and artificial intelligence
2018 · 7.493 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 · 5.438 Zit.