OpenAlex · Aktualisierung stündlich · Letzte Aktualisierung: 20.03.2026, 20:57

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

2023·7 Zitationen·Research Square (Research Square)Open Access
Volltext beim Verlag öffnen

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

Autoren

Institutionen

Themen

Artificial Intelligence in Healthcare and Education
Volltext beim Verlag öffnen