environmental harm caused by AI, including calls for transparency regarding energy consumption in AI development and operation, the use of renewable energy, and restraint in the unchecked construction of data centers. However, it may not be easy for users to intervene in or influence the environmental impacts of AI providers from a user’s position. Nevertheless, it remains important to continue exploring and pursuing practical actions that we can take within our own scope of influence. First, efforts should be made to reduce unnecessary use of generative AI. While it is not always clear what constitutes the minimum necessary level of use, environmental impacts should be kept in mind whenever generative AI is employed. For example, users should avoid unnecessary interactions such as exchanging courtesy messages with chatbots, and, in particular, refrain from generating images, audio, or video—tasks that consume far more energy than text generation—unless they are genuinely needed. If the same requests arise repeatedly within an organization, unnecessary prompts can be reduced by reusing previously generated outputs or sharing results among staff members. At the same time, care must be taken to verify the timeliness and accuracy of stored materials and to prevent inappropriate sharing of personal data across teams during internal sharing processes. For tasks that can be handled without generative AI, appropriate alternative tools—such as conventional search engines or offline data analysis tools—should be prioritized. Where possible, lighter-weight AI models (for example, ChatGPT 4o 92 93