The Discourse of Artificial Intelligence in Social Media
2023 (English)Independent thesis Advanced level (degree of Master (One Year)), 40 credits / 60 HE credits
Student thesis
Abstract [en]
Artificial Intelligence has demonstrated the last years significant tangible advantages that have been integrated in several modern services and products.Its rapid advance has generated several discussions in social media.This work provides insights into the sentiments of these discussions and identifies the major topics related to the polarizing edges of these discussions.The problem is of interest as discussions in social media provide an understanding of how the general public perceives such a disruptive technology, as well as it identifies aspects that are a public wish or concern and can proactively address it.The approach collected and analyzed a dataset of 852401 tweets with machine learning based techniques, both for sentiment analysis and topic analysis on the positive and negative tweets and also shed more light on the relations among the identified topics.With the insights acquired, several suggestions have been made that are of interest to various stakeholders and also future research avenues have been proposed.The results show overall that positive tweets are three times more than the negative ones, and that certain topics are highly controversial as they appear both in the negative and positive tweet group.In addition to understanding the expectations about potential positive impact of AI, equally important is the identification of topics that concern the citizens, and which can be further addressed with a better understanding of technology as well as safeguard measures in order to guarantee that AI as a disruptive technology will be used for the societal benefit and not to amplify potential biases or misuse that enables societal discrimination.
Place, publisher, year, edition, pages
2023. , p. 48
Keywords [en]
Artificial Intelligence, Generative AI, Natural Language Processing, Sentiment Analysis, Topic Modeling, Social Media, Twitter, Transformers, BERT
National Category
Language Technology (Computational Linguistics)
Identifiers
URN: urn:nbn:se:lnu:diva-123895OAI: oai:DiVA.org:lnu-123895DiVA, id: diva2:1791296
Subject / course
Digital humanities
Educational program
Digital humanities, master programme, 60 hp
Supervisors
Examiners
2023-09-112023-08-242023-09-11Bibliographically approved