Two stories landed this week that look unrelated but are basically the same argument: at the Met's Costume Institute, a dispute has erupted between artist Anouska Samms and designer Yoav Hadari over who gets credited for a hair dress in the exhibition. Simultaneously, Anthropic's new Claude Design tool is letting brand teams generate and iterate visual outputs through natural-language prompts. The question being fought out in both rooms is the same: when a creative object emerges from a system of relationships, who gets the byline?
Collaboration as Legal Ambiguity
The Samms-Hadari case is a classic collaborator credit dispute, the kind that used to get resolved by a handshake or a contract. But the Met's framing, that this is a dispute between the two parties rather than a failure of institutional due diligence, is revealing. Museums have always been in the business of assigning singular authorship to collective labor. A 2021 paper in the Journal of Cultural Economics by Graddy and Szymanski found that attributed single authorship in auction markets commands a measurable price premium, which means the credit question is never purely symbolic. There is money at stake in the byline.
Claude Design and the New Ghost in the Machine
Claude Design makes the same problem structural rather than interpersonal. When a brand team uses natural-language prompts to generate typography and visual systems, the labor is distributed between the prompter, the model, the training data's original creators, and the engineers who built the pipeline. TurboFund's Signal Report shows AI/LLMs leading all sectors this week with 6 high-conviction investor signals, which means capital is moving fast into this space before the authorship frameworks catch up. The Samms-Hadari case offers a preview of what happens when those frameworks don't exist: public dispute, institutional embarrassment, and the work itself getting overshadowed by its provenance fight. The irony is that Anthropic, which is simultaneously courting small business owners with AI tools, is handing collaborative ambiguity to exactly the kinds of operators who are least equipped to contractualize it.