A news summary of the article "Buying AI: five questions that should shape the contract" states that AI contracting discussions should move beyond familiar issues like data use, IP risk, and output ownership. The article proposes five starting questions. First, the role of the AI system should shape the contractual focus, as an internal tool has a different risk profile from a customer-facing or regulated system. Second, buyers must understand where the supplier adds value, as suppliers may use off-the-shelf models with value in prompt engineering or retrieval-augmented generation, meaning a clause on model training may miss how the service works. Third, the contract and governance must be balanced; governance controls like user training and human oversight should support contractual protections, but governance should not substitute for supplier accountability. Fourth, as AI procurement moves beyond standalone tools, agentic AI and MCP-style architectures mean a system may connect to other tools and data sources.