Source: JD Supra

Following the Supreme Court’s denial of certiorari in Thaler v. Perlmutter, creators lack clear federal guidance on the human-authorship requirement for copyright protection. States have started to address this gap: Arkansas enacted a 2025 statute creating an ownership right—distinct from federal IP rights—for outputs of generative AI, assigning ownership to the person who provides the input or directive to the generative AI tool, subject to a noninfringement condition and an employment exception treating certain works as akin to “work for hire.” The Arkansas law also states that a person who lawfully provides data or input to train a generative AI model shall own the resulting trained model, provided the data was lawfully acquired and ownership was not transferred, while leaving unspecified what rights or remedies apply to enforce these ownership interests.

Federal preemption under 17 U.S.C. §301(a) poses a potential barrier because courts preempt state law that governs subject matter within copyright and that confers rights equivalent to federal exclusive rights; application of the two-pronged preemption test to AI-generated works—especially where fully AI-generated works have been excluded from federal copyright protection—remains uncertain. Iowa lawmakers filed companion bills in 2026 that closely track Arkansas’s language but neither bill left committee before the session ended; the piece notes that state laws governing AI outputs could face preemption challenges, that inconsistent standards across states may produce cross-jurisdictional confusion for creators and businesses, and that pending federal litigation such as Allen v. Perlmutter could ultimately clarify the degree of human involvement necessary for copyright protection. The update emphasizes the continued need for a coherent federal framework while advising stakeholders to monitor developments at both state and federal levels.