This thesis examines the societal response to ChatGPT, a cutting-edge AI chatbot launched by OpenAI. Using Everett Rogers’ Diffusion of Innovations theory and Parasuraman’s tech- nology readiness framework, we explore how information about ChatGPT spreads through tweets and how people perceive ChatGPT. We analyze the narrative’s evolution over time, the sentiments expressed, and prevalent discussion topics. The survey was conducted to measure technology readiness for ChatGPT among its current and potential users. 325 respondents participated by filling out the online questionnaire. The research is split in three parts - volume analysis, sentiment analysis and topic modeling, and exploring technology readiness survey re- sults. Volume analysis of tweets revealed a dynamic lifecycle of public discussion of ChatGPT in Twitter resembling the Guseo-Guidolin model, with peaks tied to significant updates. Sen- timent analysis, employing VADER with 64.6% accuracy on a labeled dataset, results in distri- bution with positive (52%), neutral (31%) and negative tweets (17%). Topics extracted through Latent Dirichlet Allocation and BERTopic provide detailed insights about ChatGPT’s poten- tial to replace Google, to create various content, and its transformative impacts on education, healthcare and business. The TRI survey validated the methodology, uncovering differences in optimism and innovativeness among current and potential users, divided by cluster analysis into 5 segments: deniers, skeptics, explorers, avoiders, and pragmatists. The results of this study provide insights into the adoption and perception of ChatGPT, contributing to the ongoing AI and human-computer interaction discourse.

Diffusion of Innovations and Technology Readiness: Analyzing public discussion on ChatGPT

GERMAN, ELIZAVETA
2022/2023

Abstract

This thesis examines the societal response to ChatGPT, a cutting-edge AI chatbot launched by OpenAI. Using Everett Rogers’ Diffusion of Innovations theory and Parasuraman’s tech- nology readiness framework, we explore how information about ChatGPT spreads through tweets and how people perceive ChatGPT. We analyze the narrative’s evolution over time, the sentiments expressed, and prevalent discussion topics. The survey was conducted to measure technology readiness for ChatGPT among its current and potential users. 325 respondents participated by filling out the online questionnaire. The research is split in three parts - volume analysis, sentiment analysis and topic modeling, and exploring technology readiness survey re- sults. Volume analysis of tweets revealed a dynamic lifecycle of public discussion of ChatGPT in Twitter resembling the Guseo-Guidolin model, with peaks tied to significant updates. Sen- timent analysis, employing VADER with 64.6% accuracy on a labeled dataset, results in distri- bution with positive (52%), neutral (31%) and negative tweets (17%). Topics extracted through Latent Dirichlet Allocation and BERTopic provide detailed insights about ChatGPT’s poten- tial to replace Google, to create various content, and its transformative impacts on education, healthcare and business. The TRI survey validated the methodology, uncovering differences in optimism and innovativeness among current and potential users, divided by cluster analysis into 5 segments: deniers, skeptics, explorers, avoiders, and pragmatists. The results of this study provide insights into the adoption and perception of ChatGPT, contributing to the ongoing AI and human-computer interaction discourse.
2022
Diffusion of Innovations and Technology Readiness: Analyzing public discussion on ChatGPT
Innovation diffusion
ChatGPT
Technology Readiness
User Acceptance
File in questo prodotto:
File Dimensione Formato  
German_thesis_ft.pdf

accesso riservato

Dimensione 10.24 MB
Formato Adobe PDF
10.24 MB Adobe PDF

The text of this website © Università degli studi di Padova. Full Text are published under a non-exclusive license. Metadata are under a CC0 License

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/20.500.12608/61384