United States — AI × Entertainment Regulation¶
⚠️ Not legal advice. → Comparison by country / Cross-cutting overview (Japanese) Japanese edition: ../../../ethics-legal/by-country/us.md
Training (fair use)¶
| Case | Holding | Key point |
|---|---|---|
| Thomson Reuters v. ROSS (Feb 2025, district court) | Not fair use | Training for a competing legal research service. Argued before the Third Circuit in June 2026; decision pending (likely the first federal appellate ruling on AI training) |
| Bartz v. Anthropic (Jun 2025, district court) | Training on lawfully acquired books is fair use | A permanent library of pirated copies is not → settled for $1.5B |
| Kadrey v. Meta (Jun 2025, district court) | Fair use (plaintiffs failed to prove harm) | Stressed the importance of market effects |
- 💡 The issue is less "training itself" than "how the data was obtained" and "market substitution"
- Entertainment companies: suits against image generators (Disney & Universal v. Midjourney, June 2025) run alongside licensing deals (Disney × OpenAI, December 2025) (→ cases F03, F04)
- Music: litigation, settlements, and licenses between major labels and Suno/Udio (→ music, Japanese)
Voice & likeness¶
- State laws: e.g., Tennessee's ELVIS Act (effective July 2024; also protects "simulations" of voice)
- Federal: the NO FAKES Act (S. 4591) cleared the Senate Judiciary Committee on June 18, 2026. Enactment uncertain
- Collective agreements: the actors' union (SAG-AFTRA, 2026) and writers' union (WGA, 2026) attach notice, bargaining, and compensation conditions to AI use (→ film-video, Japanese)
AI companions & minors¶
| State / agency | Effective | Main duties |
|---|---|---|
| New York (General Business Law Art. 47) | Nov 5, 2025 | Notify all users that it is AI (at the start and every 3 hours); respond to expressions of self-harm or harm to others. Civil penalties up to $15,000 per day |
| California (SB 243) | Jan 1, 2026 | AI notice, crisis-response protocol, 3-hourly break reminders and sexual-content safeguards for minors, private right of action. In-game bots limited to game topics are excluded |
| Federal Trade Commission (FTC) | Sep 2025 | Ordered 7 companies to report on safety measures for children and teens (6(b) inquiry) |
Implications for entertainment businesses¶
- 💡 Whether a conversational NPC is "limited to game topics" determines whether California's law applies (→ ai-characters P01)
- For AI use of actors and voice actors, check union agreements in addition to the person's consent (→ case C08)
Sources¶
See the references (fairuse / usvoice / companion) in the cross-cutting overview (Japanese) and the sources for each case in case studies.