and provides AI tools itself, or uses types of AI tools other than generative AI, separate principles and guidelines shall be established. This policy focuses primarily on cases in which an organization uses commercial generative AI services such as ChatGPT, Gemini, or Claude. However, because the ways in which organizations use AI can vary widely, it may not be appropriate to apply this policy uniformly to all use cases. For example, situations such as deploying a chatbot on an organization’s website to provide information or respond to inquiries, or using AI-based real-time interpretation services at international conferences or events, may not lend themselves to the direct application of this policy. In the case of AI simultaneous interpretation services, even if hallucinations appear in the interpreted output, it may be difficult to respond to them in real time. That said, when considering whether to adopt such services, organizations can and should conduct a rigorous prior assessment based on the principles and guidelines set out in this policy. This policy may also be applied when an organization builds and uses its own generative AI system based on open-source models; however, in such cases, separate policies addressing the AI development process itself should be established. Finally, when using non-generative AI systems designed for specific purposes—such as AI systems for analyzing climate data or detecting online disinformation or hate speech—separate policies and guidelines tailored to those systems will be required. 64 65