Reading Public Sentiment Through AI: Communicative Meaning-Making and Literacy Dynamics of X Posts Surrounding Indonesia’s Free Nutritious Meal Program

Abstract

This article reports how public discourse surrounding Indonesia’s program makan bergizi gratis (PMBG) [free nutritious meal program] evolves through digital platforms and how AI-enabled sentiment analysis can illuminate broader communicative dynamics. Drawing on 25,000 X (formerly Twitter) posts collected during two implementation phases, the study analyzes shifts from early skepticism to more evaluative and neutral sentiment as the program matured. Rather than treating sentiment merely as a computational output, the analysis situates these patterns within communication scholarship on legitimacy, public engagement, and collective meaning-making. Negative discourse centered on distribution inequity, food quality, and delays, reflects broader concerns about institutional responsiveness, while the dominance of neutral sentiment in later stages illustrates the public’s move toward reflective assessment once concrete experiences accumulate. The study also identifies a notable absence of dialogic and educational content within official communication, limiting opportunities for participatory learning and weakening narrative coherence across stakeholder groups. From these findings, the article conceptualizes an AI-supported adaptive communication framework that integrates real-time sentiment insights with reflective, audience-centered public messaging. This contribution demonstrates how computational methods can support theoretically informed analyses of digital discourse, offering communication scholars a pathway for interrogating how public meaning is produced, contested, and transformed in data-intensive environments.