AI, Deepfakes & Digital RightsAI Fundamentals

What Is Generative AI?

Executive Overview

Generative AI refers to artificial intelligence systems that produce new content — text, images, audio, video, or code — by learning statistical patterns from large datasets of existing human-created works. Unlike rule-based software, generative AI models produce novel outputs that resemble training data. The major categories relevant to entertainment law are large language models (LLMs) for text and scripts, diffusion models for images and video, and neural audio synthesizers for music and voice. In entertainment law, generative AI is significant because the content it produces and the data used to train it implicate copyright, right of publicity, and contract law in ways courts and legislatures are only beginning to resolve.

Why It Matters

Generative AI is the defining legal disruption of the current entertainment industry. Every major guild negotiation since 2023 has centered on AI provisions. Studios, streaming platforms, and independent creators are grappling with how AI tools change the economics of content creation, who owns AI-generated output, and what protections exist for human creators whose work trained these systems. Attorneys who cannot explain how these systems work at a functional level will struggle to counsel clients effectively — whether drafting AI provisions in talent agreements, advising on training data licensing, or litigating infringement claims.

Statutory Foundations & Regulatory Framework
17 U.S.C. § 102(a)

Establishes that copyright protection extends to original works of authorship fixed in a tangible medium — the human authorship requirement that the Copyright Office applies when evaluating AI-generated content for registration.

17 U.S.C. § 107

The fair use doctrine — the primary defense asserted by AI companies arguing that training on copyrighted works without a license is permissible.

NO FAKES Act of 2026 (S. 4591 — pending)

Would create a federal right giving every individual control over digital replicas of their voice and visual likeness, directly targeting AI-generated synthetic performers. Ordered reported by the Senate Judiciary Committee on June 18, 2026.

SAG-AFTRA 2023 TV/Theatrical Contract — AI Rider

Requires informed written consent, daily-rate compensation, and approval rights for any digital replica of a performer created or used in a production.

WGA MBA 2023 — Article 72

Prohibits studios from using AI to generate material that would otherwise require a WGA writer, and from requiring writers to use AI tools. Requires disclosure when AI-generated material is provided to a writer.

Major Cases
Andersen v. Stability AI Ltd.N.D. Cal. 2024
Legal Issue

Whether AI image generators trained on artists' copyrighted works without license infringed those works, and whether the four-factor fair use test applied to AI training at scale.

Holding & Impact

Direct copyright infringement claims survived early dismissal. The court rejected Stability AI's argument that training on copyrighted images was categorically non-infringing. Trial set for 2026 — will be the first major AI training data case to go to trial.

Read full opinion →
RIAA v. Suno Inc.D. Md. 2024
Legal Issue

Whether an AI music generator trained on copyrighted sound recordings without license infringed those recordings.

Holding & Impact

Landmark industry lawsuit filed by major labels against one of the leading AI music generators. Settlement reached in 2025 for undisclosed terms — establishes that unauthorized training on commercial recordings carries significant infringement exposure.

Read full opinion →
Landmark Case
Thaler v. PerlmutterD.D.C. 2023
Legal Issue

Whether a work generated entirely by AI without human authorship qualifies for copyright protection.

Holding & Impact

The court affirmed the Copyright Office's refusal to register a work with no human author, establishing that current U.S. copyright law requires human authorship as a baseline condition for protection.

Read full opinion →
Industry Impact

The entertainment industry sits at the epicenter of generative AI disruption because it produces exactly the high-quality, expressive content AI systems need for training. Studios face pressure to deploy AI tools to reduce production costs while simultaneously negotiating with guilds that have made AI restrictions central bargaining demands. The 2023 strikes demonstrated that labor can impose meaningful contractual limits on AI — but enforcement and compliance monitoring remain largely untested. Independent creators face the sharpest asymmetry: their works are used to train systems they cannot afford to litigate against, and the guild protections that cover major studio productions don't apply to them.

Practical Tips
01

Before advising any entertainment client on AI use, understand the specific AI tools being used — LLMs, image generators, and voice synthesizers have different legal risk profiles.

02

Every production agreement involving any AI tool should address: (1) which tools are permitted, (2) who owns AI-generated output, (3) what disclosure obligations apply, and (4) guild compliance.

03

For talent clients, negotiate AI provisions before signing — retroactive renegotiation is significantly harder once a studio has established practices.

04

Track the NO FAKES Act progress closely — if enacted, it will be the most significant change to right of publicity law in decades and will affect virtually every entertainment contract involving talent.

05

Advise label and publishing clients to register copyrights in their catalogs if they haven't — registration is a prerequisite to statutory damages in infringement suits against AI companies.

Key Takeaways
01

Generative AI produces novel outputs by learning patterns from training data — the training process itself is the central copyright dispute.

02

AI-generated works without human authorship cannot be copyrighted under current U.S. law.

03

Guild agreements (WGA, SAG-AFTRA) now impose disclosure, consent, and compensation requirements for AI use in covered productions.

04

The NO FAKES Act, if passed, would create the first federal right over AI-generated voice and likeness replicas.

05

Entertainment attorneys need functional knowledge of how AI systems work to draft effective provisions and advise on liability exposure.

FAQs
Can AI-generated content be copyrighted?

Not if generated entirely by AI without sufficient human creative control. The U.S. Copyright Office will register works where a human author made meaningful creative choices in the process, but the AI-generated elements themselves receive no protection.

Is it legal to train an AI on copyrighted entertainment content?

Legally unresolved. AI companies assert fair use; rights holders assert infringement. No appellate court has issued a definitive ruling on AI training specifically. Multiple major cases are pending, and the outcome will define the industry for years.

Do guild agreements restrict AI use in productions?

Yes, significantly. The 2023 WGA and SAG-AFTRA agreements require consent, disclosure, and compensation when AI tools are used to generate or modify covered content, and prohibit studios from using AI to replace covered work without guild approval.

How do I know if an AI tool creates infringement exposure for my client?

Assess the tool's training data (was it trained on licensed or unlicensed content?), the output (does it reproduce recognizable elements of specific works?), and the commercial context. Tools trained on licensed data with output restrictions carry lower risk than those trained on scraped content.

Resources & External Links