ChatGPT or Gemini: Who Makes the Better Scientific Writing Assistant?

Abstract

The rapid evolution of scientific research has created a pressing need for efficient and versatile tools to aid researchers. While using artificial intelligence (AI) to write scientific articles is unethical and unreliable due to the potential for inaccuracy, AI can be a valuable tool for assisting with other aspects of research, such as language editing, reference formatting, and journal finding. Two of the latest AI-driven assistants that have become indispensable assets to scientists are ChatGPT and Gemini (Bard). These assistants offer comprehensive support from literature review to journal suggestion, and they have the potential to revolutionize research. In the present study, a comprehensive set of queries and responses were designed to assess the capabilities of ChatGPT-3.5 and Gemini as scientific assistants. The results showed that Gemini achieved a perfect score of 100%, while ChatGPT-3.5 scored a less impressive 70%. Notably, ChatGP-3.5 fell short in specific areas that includes providing assistance with scientific paper explanations, exploring bibliographic databases, and formatting references. The qualitative assessment of responses also suggests that both the AI chatbots can be valuable tools for researchers, however, Gemini seems to be more appealing and accurate through the whole procedure of scientific writing. This work shall open new research dimensions in identifying adequate scientific utilization of the evolving AI tools and technologies, as embracing these advancements will be essential for staying at the forefront of scientific research.