Case Studies
Deep dives into specific decisions and determinations shaping AI copyright analysis
These case studies go beyond summary mentions in the legal framework pages. Each study focuses on procedural posture, holding scope, what the outcome does not decide, and concrete implications for developers and maintainers.
Featured Studies
Thaler v. Perlmutter (US, D.C. Circuit, 2025)
Topic: Human authorship boundary.
Clarifies that wholly autonomous AI output cannot be registered with AI as sole author, while explicitly leaving AI-assisted human-authored works outside the holding.
Zarya of the Dawn (USCO, 2023)
Topic: Partial registration model.
Shows how text and arrangement can be protected while individual AI-generated images are disclaimed.
Theatre D'opera Spatial (USCO Review Board, 2023)
Topic: Claim framing and disclaimer strategy.
A key example of denial tied to what was claimed, not a blanket rule against all AI-assisted creative work.
Invoke: A Single Piece of American Cheese (USCO record, 2025)
Topic: Documentation-heavy success path.
Illustrates how iterative edits, selection, and provenance improve registration posture.
EU AI Act / Copyright Enforcement (EU, 2024-2026)
Topic: Training-data/TDM legality in EU context.
Complements output-authorship analysis by focusing on training-phase legal constraints, opt-outs, and Article 50 transparency obligations.
Japan Context (How It Maps to These Cases)
Japan currently has fewer AI copyright court decisions than the US, but its statutory and policy structure is highly relevant when interpreting these cases in global operations:
- Training phase: Article 30-4 is comparatively permissive for non-enjoyment purpose data use.
- Output phase: Standard infringement and similarity/dependence analysis still applies.
- Authorship: Human creative contribution remains required for output-level protection.
See the Japan framework analysis in Japan Legal Framework (EN/日本語), compare with EU Legal Framework, and use Practical Guide for documentation steps.