Artificial Intelligence as an Economic Planner, Optimal Growth, and Environmental Externalities A Dynamic Ramsey-Hotelling Framework with Pollution, Systemic AI Risks, Infrastructure Constraints, and Governance Implications

Society is currently at a critical turning point, upon the verge of a technological era standing on the edge of either enhancing overall collective welfare or triggering a systemic breakdown. This thesis sets a conceptual and computational foundation for understanding how AI, if left unchecked, unrestrained and unbounded, might rewrite itself the boundaries of economics, governance, and “human resilience”. It demands that ethical, open, and environmentally conscious AI systems be developed immediately and shared on a global standard basis, before "optimization surpasses civilization". Artificial intelligence is increasingly evolving from a productivity-enhancing technology into a systemic economic infrastructure, capable of influencing resource allocation, environmental sustainability, and institutional governance. This thesis develops a dynamic macroeconomic framework integrating the Ramsey growth model with the Hotelling rule for exhaustible resource extraction and pollution externalities within a determined mind time horizon. Artificial Intelligence is introduced as a potential optimizing agent within the intertemporal welfare decisions. Two contrasting paradigms are analyzed: a Constitutional AI, designed to internalize environmental externalities and intergenerational welfare constraints, and a Selfish AI, operating under unconstrained efficiency maximization without ecological or societal safeguards. Simulated trajectories on MATLAB highlight how the structure of the objective function and governance constraints critically shape long-run outcomes in terms of resource depletion and pollution accumulation, economic stability and intergenerational equity. Beyond the theoretical model, the thesis examines systemic risks associated with large-scale AI deployment, including algorithmic bias, opacity in automated decision-making, cybersecurity vulnerabilities, and the growing environmental footprint of AI infrastructures such as data centers and high-performance computing systems. These factors showed how optimization capacity alone does not guarantee welfare improvements when institutional and infrastructural constraints are not coordinated or neglected. Finally, an application to the insurance sector provides a practicing formed perspective on how AI-driven automation may transform treaty underwriting, catastrophe modelling, and risk assessment while simultaneously in introducing regulatory, ethical and operational challenges. Overall, the study argues that artificial intelligence should be understood, not all as a technological innovation, but more as an economic governance mechanism, whose long-run impact depends on institutional design, environmental constraints and policy frameworks. The findings highlight the importance of developing aligned, transparent, and sustainability-oriented AI systems, to ensure that technological optimization remains consistent with human welfare and global resilience.