Mapping Market Risk Narratives of Artificial Intelligence: A Social Network Analysis of Public Discourse

Abstract

As AI technologies rapidly progress, privacy and surveillance, cybersecurity, and governance have become increasingly pressing concerns that raise questions about how the narratives of risk and AI influence the market's perception of new technologies. As a component of the public discourse, it is becoming more and more a sign of trust or distrust in technology and a source of legitimation in the digital world. This study utilized a Social Network Analysis (SNA) and Text based analytical method to analyze the perception of the market on AI by analyzing the AI risk narratives in the online public discourse. Data were gathered from X (formerly Twitter) using the SocialX platform over the first quarter of 2026 (1st January-31st March 2026) and found to be 2,883 public posts. The analysis combined trend analysis, sentiment analysis, temporal word analysis, text network analysis, BERTopic topic modeling, zero-shot classification, and SNA. The results showed an attitude of caution, with neutral (57.86%) and negative sentiment (35.93%) dominating the AI risk discourse, suggesting that the public isn't entirely positive about AI. Most prevalent data privacy issues were Data Privacy Risk (43.81%), Surveillance and Control Risk (21.19%) and Cybersecurity Risk (11.24%). Text network analysis further identified data, training, and surveillance as central themes within the discourse. This study contributes theoretically by integrating Framing Theory, Perceived Risk Theory, and market legitimacy perspectives to explain AI risk narratives as indicators of market perception. In practical terms, the results offer insights for trust-building in AI, responsible communication, and technology branding strategies.