Sentiment Analysis on Generative Large Language Models based on Social Media Commentary of Industry Participants

Oana-Alexandra Miron & Khant Nyein Htet Wai

Student thesis: Master thesis


OpenAI released their new product, ChatGPT on the 30th of November 2022. ChatGPT is a generative large language AI model and an extension of GPT-3.5, that makes the interaction between the user and the machine more accessible to the general public. The AI-powered chatbot responds to user requests with almost real-time dialogue in a naturally spoken way. The use cases of ChatGPT for businesses are long discussed across media and within business environments. These are manifold and include creating source code, producing creative texts, and assisting in repetitive tasks. With such promising prospects, ChatGPT managed to reach a 100 million user base in record time. Thus, the purpose of this paper is to explore which industries are more open to accepting the technology of generative large language models based on opinions about ChatGPT. Since Twitter became a public forum where well-known figures and regular people are expressing their opinions alike, this paper conducts topic modeling and sentiment analysis on tweets related to ChatGPT. The aim is to map the authors of these tweets to industries and to subsequently uncover the prospects of acceptance of the technology within these different industries. The research finds that the people expressing their opinion about ChatGPT on Twitter either do not relate their profiles to an industry or are part of industries such as: Education & Academia, Entrepreneurship & Business Leadership, Finance & Crypto, and more. While tweets vary from industry to industry and from user to user, this research showcases that the general sentiment over generative large language models is a positive one, which is then discussed from the perspective of technology acceptance and its implications. This paper contributes to technology acceptance research, exhibiting the use of Twitter data to explore perception and acceptance in relation to generative large language models.

EducationsMSc in Business Administration and E-business, (Graduate Programme) Final Thesis
Publication date2023
Number of pages103
SupervisorsAbid Hussain