A live intelligence layer tracking emerging copyright class actions, model training disputes, and regulatory rulemakings — calculating the exact doctrinal fault lines each outcome will rewrite.
Select an active frontier lawsuit and toggle hypothetical judicial rulings to simulate the domino effect across the master legal graph.
“Does mass training on millions of paywalled articles without a license constitute commercial copyright infringement under post-Warhol Factor 1?”
Training LLMs on copyrighted expressive prose for synthetic commercial output is not transformative fair use; market substitution under Factor 4 is dispositive.
U.S. District Court for the Southern District of New York (S.D.N.Y.) · Docket 1:23-cv-11195
The New York Times alleges OpenAI and Microsoft engaged in massive unauthorized copying of millions of copyrighted news articles to train GPT-4, producing synthetic outputs that reproduce verbatim and near-verbatim text and usurp Search/Browse markets.
If GPT-based search tools regurgitate full journalistic paragraphs, it directly destroys digital subscription and syndication markets under Warhol and Factor 4.
Tests whether LLM training is non-expressive computational mining (Authors Guild) or an unauthorized derivative authoring engine (Warhol).
Claims that stripping bylines, titles, and terms of service during dataset tokenization violates § 1202(b).
U.S. District Court for the Northern District of California (N.D. Cal.) · Docket 3:23-cv-00201
Visual artists class action claiming that text-to-image latent diffusion models (Stable Diffusion, Midjourney) ingest billions of copyrighted artworks from LAION-5B to compress them into mathematical parameters, functioning as continuous derivative-work generation engines.
Courts must rule whether statistical model weights storing mathematical associations constitute unauthorized 'derivative works' under 17 U.S.C. § 106(2).
Judge Orrick sustained direct copyright infringement claims against Stability AI for local image caching during model training.
Addresses whether invoking artist names in prompt templates without permission constitutes false endorsement or trademark violation.
U.S. District Court for the Middle District of Tennessee (transferred to N.D. Cal.) · Docket 3:23-cv-00444
Major music publishers sue Anthropic, alleging Claude language models systematically ingested and reproduce verbatim copyrighted song lyrics upon prompt requests, bypassing licensing agreements and mechanical royalties.
Lyrics constitute pure expressive text with zero functional code defense, challenging Anthropic's general fair use defense.
Argues Anthropic's engineering of system prompts and guardrails proves knowledge and control over infringing output generation.
Library of Congress / U.S. Copyright Office · Docket Docket No. 2023-6
Comprehensive federal study and regulatory guidance examining (1) the use of copyrighted works to train AI models, (2) the copyrightability of AI-generated outputs, (3) potential statutory licensing frameworks, and (4) federal right of publicity legislation.
Maintains strict rule that solely machine-generated prompts cannot receive copyright registration without substantial human expressive control.
Recommends whether Congress should intervene with statutory compulsory licenses (similar to mechanical radio royalties) for AI training data.
N.D. Cal. (Judge Rita F. Lin) · Docket No. 4:23-cv-00770
Nationwide class action testing whether AI software vendors can be held directly liable under Title VII, ADEA, and ADA as statutory 'employment agencies' when enterprise clients delegate automated candidate screening and ranking discretion.
Judge Lin denied Workday's motion to dismiss, holding that AI vendors exercising delegated hiring discretion are subject to direct federal civil rights liability.
Tests whether statistical disparities in machine learning model outputs establish prima facie discrimination under Griggs.
National Labor Relations Board / Administrative Law Judge · Docket NLRB Case No. 29-CA-280153
NLRB administrative proceeding challenging automated 'Time Off Task' (TOT) algorithmic tracking and automated employee termination algorithms as unlawful interference with Section 7 concerted labor activity.
NLRB General Counsel memo GC 23-02 establishes that omnipresent algorithmic tracking presumptively chills union organizing and concerted employee discussions.
Evaluates whether automated deactivations and firing without human managerial discretion violate statutory duty of fair representation.