The Impact of Artificial Intelligence on Post-Keynesian Stock-Flow Consistent (SFC) Modeling and Accounting

This chapter examines how Artificial Intelligence (AI) can enhance Post-Keynesian Stock-Flow Consistent (SFC) models by improving the modeling of confidence, a traditionally elusive but central variable. Using the Tunisian economy as a case study, the paper builds on a six-sector SFC model and proposes a dynamic confidence index (CI) derived from AI methods, including natural language processing and deep learning. By integrating structured and unstructured data, such as media sentiment, tourist flows, and economic indicators, the AI-driven CI captures real-time shifts in collective expectations. The result is a model that reacts faster to shocks (e.g., revolutions, pandemics), better simulates behavioral effects, and strengthens forecasting performance. This integration does not replace but complements the theoretical foundations of SFC modeling, making it more responsive to contemporary economic complexity. The study demonstrates that AI can transform confidence into an endogenous, learning-driven variable, enhancing both realism and policy relevance in macroeconomic simulations.