Exploring the Reactions of Early Users of ChatGPT to the Tool using Twitter Data: Sentiment and Topic Analyses
Achraf Tounsi, Safa Elkefi, Safa Bhar Layeb
Abstract
Recently, large language models have gained considerable interest due to their impressive performance on different tasks. OpenAl's ChatGPT is a revolutionary language model that has quickly become popular among early adopters. Many users have gone so far as to call it a disruptive technology in many different domains. Its large, pre-trained model provides powerful capabilities already being applied to various use cases. Gaining an understanding of the thoughts and feelings of those who are quick to embrace new technologies is essential, as it can offer valuable insights into the technology's potential successes, failures, and other characteristics. In this paper, a mixed-method study was developed using 463,983 tweets from early ChatGPT users. We first apply topic modeling to identify the maj or topics and then conduct an extensive qualitative sentiment analysis. We observed that most people were pleased with the tool, but some expressed apprehension about its potential ethical and legal implications. After its launch, however, the fears were largely alleviated as people began to use it and recognize its benefits for personal and business purposes. It is important to be aware of any new technology's possible drawbacks and use it responsibly.