<![CDATA[Background: Central-bank communication increasingly governs expectations through forecasts, conditional projections, and policy signals, while artificial intelligence expands the capacity to classify and compare such language across heterogeneous monetary regimes. Objective: This study examines how central banks in Indonesia, Singapore, Malaysia, and the Philippines construct economic futures, authorise policy judgement, and become analytically legible through AI-assisted textual modelling. Method: A comparative corpus-assisted discourse design analyses fourteen official English-language policy communications using predictive-language annotation, communicative-power coding, normalised institutional profiles, cosine similarity, concordance checking, and human validation. Results: Predictive discourse consistently combined epistemic modality, temporal projection, and inflation alignment, although quantification, conditionality, directional risk, and forecast revision varied across institutions. Communicative authority emerged through different configurations: Bank Indonesia foregrounded policy commitment and exchange-rate stability, MAS emphasised numerical forecasting and recalibration, BNM contextualised projections through broader macroeconomic conditions, and BSP combined formal decisions with forecast monitoring. Implication: Computational comparison identified substantial but incomplete institutional convergence, demonstrating that shared monetary vocabulary did not erase differences in mandate, genre, or policy orientation. Novelty: This study contributes an integrated account of predictive authority by linking linguistic futurity, institutional performativity, and explainable AI, while treating computational outputs as interpretive evidence requiring contextual and human scrutiny across linguistically and institutionally differentiated Southeast Asian monetary systems]]>
Copyrights © 2026
