1. Investigating the Prospects of ChatGPT in Training Medicinal Chemists and the Development of Novel Drugs
- Author
-
Michell O. Almeida, Artur C. G. Soares, Gustavo H. M. Sousa, Witor R. Ferraz, and Gustavo Trossini
- Subjects
Natural Language Processing ,Medicinal Chemistry ,Drug design and development ,Chemoinformatic ,Computational chemistry ,Science ,Chemistry ,QD1-999 - Abstract
This scientific article delves into the advantages, insights, and limitations of ChatGPT in various scientific domains. Alongside other large language models, this tool exhibits the potential to directly or indirectly assist in a range of scientific areas including Computer Science, Chemistry, Biology/Bioinformatics, and Medicine. Some of the functionalities of ChatGPT include text translation, code improvement, data visualization, and database cleaning. The model can aid in writing and translating scientific articles from mostly any language. In the field of chemoinformatics and computational chemistry, ChatGPT can provide code examples and assist in code development, by evaluating and enhancing code readability and project documentation. Furthermore, it can assist in the database cleaning process and create customized functions for performing specialized tasks. However, ChatGPT does possess some limitations, such as frequent occurrences of artificial hallucinations (a response generated by AI that comprises erroneous or misleading information presented as true), the inability to process multimodal information, and the potential for biases in its training datasets. Therefore, caution must be exercised when incorporating these technologies, considering their social impact and implications for the job market. Acknowledging limitations is crucial when using these tools. With careful and proper use, they can aid the scientific process with the potential to speed up drug discovery.
- Published
- 2024
- Full Text
- View/download PDF