Sentiment Analysis of Research on AI Ethics: A Web-Based Study

The field of Artificial Intelligence has experienced significant growth in recent years, both in terms of technological development and global adoption rates. AI-based solutions are now finding their way into the lives of an increasing number of people worldwide, being used both professionally and personally. However, along with this progress, numerous controversies have arisen regarding the ethics of their use in various domains. In the specialised literature, there are a growing number of publications focused on the discussion of this topic. The present study focusses on analysing the general trends of these scientific works in relation to the ethics of the use of AI. Using the Web of Science Clarivate database, a set of publications was selected based on keywords and subsequently subjected to analysis. Sentiment analysis techniques are used to identify the positive or negative trend among specialists and how it has evolved over the years. Latent Dirichlet Allocation is used to highlight the main topics developed in these writings. On the basis of the obtained results, it can be noted that the concern for AI ethics issues is increasingly addressed in specialised writings. Sentiment analysis reveals that, in recent publications, on average, sentiments tend to be slightly positive, but the polarity value has decreased in recent years. Thus, this study contributes to a better understanding of specialists' positions regarding identified AI ethics issues, highlighting results obtained through the application of modern natural language processing techniques and by presenting important aspects emphasised in existing scientific works on this topic.