The billboard showed a burning building. The dog was still inside, still sitting at a table, still saying "This is fine." But this time, the meme that KC Green drew in 2013 had been repurposed—not by a meme account, not by a fellow internet humorist, but by an AI startup called Artisan, whose other billboards urged businesses to "stop hiring humans." Green is now suing, alleging his work was used without permission. The question his lawsuit raises is deceptively simple: when your art becomes the internet's most recognizable symbol of denial, who gets to say what happens to it next?
The case is the latest flashpoint in a fight that has been simmering since generative AI became a consumer product. Artisan scraped and trained on vast quantities of online imagery—including, according to Green's complaint, the "This is fine" comic—to build its models. Then it used that cultural currency to sell its product in high-profile ad campaigns. Green argues that neither the training nor the marketing was authorized. The startup has not publicly commented on whether it trained on Green's work or merely used it in advertising, but the distinction may not matter legally if a court determines that both constitute infringement.
This is where the case gets interesting for the broader AI industry. Artisan is not the first company to find itself accused of treating the open internet as a free resource.训练数据——the images, text, and audio used to teach AI systems—has become the industry's equivalent of oil: essential, valuable, and, so far, mostly taken without formal negotiation. Companies have argued that scraping for training purposes falls under fair use. Courts have not definitively resolved this. Green's lawsuit now asks whether using that same work in marketing—after the training is done—crosses a line that training alone might not.
The timing matters. The U.S. Copyright Office has been developing clearer guidelines around AI-generated content and training data, moving toward positions that give human creators more leverage. Internationally, the EU AI Act has introduced mandatory disclosure requirements for training data, though it stops short of mandating compensation. Green is not the only creator pursuing this path; a wave of similar lawsuits is working through courts, and their combined weight could reshape how the industry operates.
For individual artists, the stakes are immediate and personal. Green's comic was a commentary on denial and apathy—about a creature too resigned to crisis to act. That resonance is precisely why AI companies want it. But the creator sees something different in how his work has been appropriated: not just a lost licensing fee, but a loss of meaning. The dog in the burning room is supposed to be a punchline about inaction. Now it sells software that its creator argues will enable more displacement and inaction, just at a larger scale.
Artisan has not filed a formal response to the lawsuit as of this writing. The company's positioning—aggressive, anti-establishment, deliberately provocative—suggests it may frame any defense around the idea that the law has not caught up with technological reality. That argument has resonated in some courtrooms. It has failed in others. What happens next will depend on which judge hears the case, which jurisdiction applies, and whether either side decides to settle before a ruling forces the issue into the open.