Bluesky user argues AI training is transformative fair use

A Bluesky user identified as sihnrae argued on October 27, 2026, that training AI models on existing artworks constitutes transformative fair use because the process converts source images into non-reproducible model weights. The post asserts this technical transformation satisfies legal standards while characterizing current copyright frameworks as less ethical than fair use doctrines. This perspective aligns with arguments previously advanced by AI companies in pending litigation against artists and publishers. Legal experts note courts have not yet established whether weight generation qualifies as transformative under U.S. copyright law. The statement reflects ongoing polarization in intellectual property debates surrounding generative AI development. No specific lawsuit or regulatory action was cited in the post. The argument remains a personal opinion without institutional backing or judicial validation.