<p>Smart cities rely on interconnected digital services to improve efficiency, safety, and user experience across urban mobility systems. However, as these services scale, traditional manually developed smart contracts introduce significant risks, including programming errors, malicious logic insertion, and the inability to adapt to dynamic environmental conditions. To address these limitations, this research proposes a guardrail-enhanced, context-aware framework that automates smart contract generation and deployment using Generative AI and a permissioned blockchain (Multichain). In the proposed approach, smart contracts are not written or updated manually instead, real-time V2X contextual data such as traffic patterns, passenger load, and environmental conditions—are processed by a predictive AI model, which triggers dynamic prompt construction for a Generative AI engine (GPT-4). Guardrails are incorporated at both the prompt and code-validation stages to ensure that the generated smart contract logic remains secure, policy compliant, and free from malicious or unsafe behavior. The validated contract is then deployed through a Python-based virtual machine, while its logic hash and metadata are immutably recorded on Multichain streams for auditability and traceability. To evaluate the feasibility of this guardrail-driven automation process, we implement a V2X fare management use case in which fare discounts are predicted using a trained machine learning model and embedded into dynamically generated smart contracts. Experimental results demonstrate that the proposed system enables secure, adaptive, and fully automated smart contract updates, significantly reducing human intervention while maintaining high throughput, reliable contract retrieval, and strong contextual responsiveness in dynamic transportation environments. This work highlights the potential of combining Generative AI, guardrails, and blockchain to create next-generation autonomous governance mechanisms for smart city ecosystems.</p>