Analyzing 3 Million Twitter Posts: Positive or Negative?

Table of Contents

1. Introduction
2. Methodology
   1. Data Collection
   2. Analysis Process
3. Public Perception of Generative AI
   1. Interest in Generative AI by Occupation
   2. Positive and Negative Reactions towards Generative AI by Occupation
   3. Topics Tweeted about by Occupation
4. Attitudes towards AI Replacement
   1. AIexposure and Emotional Score
5. Usage of ChatGPT
   1. Positive and Negative Uses of ChatGPT by Profession
6. Change in Interest in AI since the Creation of Generative AI
   1. Number of Users Tweeting about AI
   2. Emotional Score of AI-related Tweets
   3. Analysis by Labels
7. Conclusion

Public Perception of Generative AI

Generative AI has become a topic of much interest and discussion in recent years, with varying opinions and reactions from different occupations. This article aims to Delve into the public perception of generative AI by analyzing millions of Twitter posts. We will explore the level of interest in generative AI across various professions, as well as the positive and negative attitudes towards this technology. Additionally, we will examine the topics most commonly discussed by different occupations and investigate how the fear of job replacement by AI affects the overall sentiment towards generative AI.

1. Introduction

Generative AI has undoubtedly revolutionized various industries, offering a range of benefits and possibilities. However, its rapid development has also raised concerns and apprehensions among individuals who fear losing their jobs to AI. This article will present an overview of a recently published paper that examines the public reaction to generative AI using Twitter data. By analyzing millions of tweets from January 2019 to March 2023, the study provides valuable insights into how people across different professions perceive generative AI.

2. Methodology

To understand the public's perception of generative AI, the study collected approximately 3 million English tweets from around 1 million users over a span of four years. The tweets were categorized into different types of generative AI, including chats, images, codes, and models. Only manually operated tweets were included, while automated tweets from bots were excluded from the analysis.

3. Public Perception of Generative AI

3.1 Interest in Generative AI by Occupation

The study begins by examining the level of interest in generative AI across various occupations. By comparing the number of posts related to generative AI to the overall number of daily posts, the study highlights the professions that Show a higher degree of interest in generative AI. IT professionals, such as data scientists, product managers, and software engineers, display a significant interest in generative AI. However, it is intriguing to note that even professionals in fields such as law and teaching show a considerable interest in generative AI, indicating its widespread impact.

3.2 Positive and Negative Reactions towards Generative AI by Occupation

Next, the study investigates the positive and negative attitudes towards generative AI among different occupations. Using visualizations, it becomes apparent that certain generative AI applications, such as image generation and model creation, are more positively perceived compared to chat-Based AI. Data scientists and product managers exhibit particularly positive responses towards generative AI, while illustrators and streamers/authors display more negative emotions. Notably, illustrators express a significant negative sentiment towards image-generating AI, possibly due to concerns about copyright issues.

3.3 Topics Tweeted about by Occupation

The study further explores the topics frequently discussed by individuals in different occupations concerning generative AI. Emotional scores are assigned to the tweets, with red indicating a higher level of positivity and Blue representing negativity. Data scientists show a positive sentiment across all topics, while artists and illustrators express negative feelings, mainly related to copyright concerns. This analysis provides valuable insights into the specific areas of interest or concern for professionals in relation to generative AI.

4. Attitudes towards AI Replacement

One might assume that individuals in professions susceptible to being replaced by AI would hold a negative view towards generative AI. However, the study reveals an unexpected finding: the more likely a job is to be replaced by AI, the more positive the individual's attitude towards generative AI. This surprising phenomenon suggests that individuals in occupations with high AIexposure scores are more open to embracing new technology and have a positive outlook towards generative AI.

5. Usage of ChatGPT

The study investigates how people are utilizing ChatGPT, one of the popular generative AI models. By analyzing tweets containing screenshots of ChatGPT, the study identifies different use cases and evaluates the sentiment associated with each use. Translators and lyricists display a positive sentiment, indicating the usefulness of ChatGPT in their respective fields. However, those entering Prompts related to the coronavirus or AI models show a more negative sentiment, highlighting potential limitations or challenges in using generative AI for certain topics.

6. Change in Interest in AI since the Creation of Generative AI

The study examines how the interest in AI has evolved since the creation of generative AI models like ChatGPT. By analyzing the number of users tweeting about AI and the frequency of Google searches, the study reveals a significant increase in interest in AI since the release of generative AI models. The emotional score of AI-related tweets also fluctuates over time, with positive sentiment gradually increasing towards the end of 2021. However, a recent surge in interest suggests a renewed Curiosity and enthusiasm for AI among the public.

7. Conclusion

In conclusion, the public perception of generative AI is a complex and multifaceted phenomenon. This article has highlighted key findings from a study analyzing millions of Twitter posts to gain insights into the worldwide sentiment towards generative AI. The study reveals varying levels of interest, positive and negative reactions, and specific concerns across different occupations. Furthermore, it challenges the assumption that individuals in professions at high risk of AI replacement hold negative attitudes. Overall, this research offers valuable insights into the evolving perceptions and attitudes towards generative AI and its impact on society.