Generative AI Output

Generative AI output is content, text, images, audio, video, or code, produced by an artificial intelligence system from a user prompt or input. Under current U.S. copyright doctrine, content produced by an AI system without sufficient human control over its expressive elements is not copyrightable. The legal framework governing ownership, allocation of risk, and use of generative AI output in commercial contexts continues to develop rapidly.

Authorship and copyrightability

Thaler v. Perlmutter (D.C. Cir. 2025) affirmed the U.S. Copyright Office's denial of registration for a work created entirely by an AI system. The court held that the Copyright Act of 1976 requires human authorship as a matter of statutory law. The Supreme Court denied certiorari in March 2026, leaving the D.C. Circuit's holding intact. The Copyright Office's January 2025 guidance further clarified that prompt engineering alone, even highly detailed iterative prompting, does not provide the human creative control necessary to establish authorship of the AI output.

Hybrid works

Works that combine AI-generated portions with substantial human authorship remain protectable, but only the human-authored portions are protected. A novel written by a human that incorporates AI-generated illustrations would be copyrightable as to the text but not as to the illustrations. The Copyright Office requires applicants to disclose and exclude more than de minimis AI-generated material. An applicant who fails to do so risks cancellation of the registration, and a court may disregard the registration under 17 U.S.C. § 411(b) if the applicant knowingly included inaccurate information that would have caused the Office to refuse it.

Training data and output infringement risk

Pending litigation against major AI developers raises distinct infringement questions concerning (1) the use of copyrighted material in training datasets and (2) AI outputs that closely resemble specific copyrighted works in the training data. These questions remain largely unsettled. Indemnification provisions in AI vendor contracts are one way businesses allocate this risk.

Texas TRAIGA framework

The Texas Responsible Artificial Intelligence Governance Act took effect January 1, 2026. TRAIGA imposes obligations on persons and entities developing or deploying AI in Texas, amends the TDPSA to clarify processor obligations for AI-handled data, and clarifies application of the Texas Capture or Use of Biometric Identifiers Act to AI training data. Enforcement authority is vested in the Texas Attorney General.