This research proposes a seven-step legal framework to determine liability among developers, deployers, and end users of AI-generated contracts. It uses doctrinal legal analysis and comparative studies of the EU, U.S., and international regulations, focusing on tools like Ironclad, Robin AI, and Legitt AI. The framework considers factors such as AI system risk level, control, contractual roles, fault, and regulatory compliance, aligning with the EU AI Act and contract law principles. It aims to allocate responsibility fairly for harms like GDPR violations, biased clauses, or unenforceable terms. The model supports transparency, fairness, and accountability while promoting responsible AI use in contract automation.