the same time, legislative debates are underway regarding whether and how the use of copyrighted works as AI training data—often referred to as text and data mining (TDM)—should be permitted or regulated. In parallel with these legislative efforts, numerous copyright lawsuits related to training data are currently in progress, and their outcomes are likely to serve as important reference points. From the perspective of users of generative AI, there is a risk of copyright infringement if the system produces outputs that are identical or substantially similar to existing works. For this reason, extra caution is required when using generative AI for publicly released content to ensure that no infringement occurs. Beyond the legal risks faced by individual users, it is also important to consider the broader context. Generative AI is not only built using copyrighted works but also competes with creative workers in the marketplace, posing economic threats to their livelihoods. To the extent that generative AI relies on structures in which creative labor is exploited without the consent or compensation of creators, it raises serious ethical and political-economic concerns. Are the data collection and content generation processes behind the generative AI tools we use transparent, and are fair compensation mechanisms in place? The Environmental Costs of Generative AI Generative AI is an environmentally expensive technology. Training large-scale models requires vast computational resources, and the carbon emissions generated in this process can amount to 34 35