Observed Signal · Aug 23, 2026 · Policy Update · Source: techcrunch · Impact: 4/5 · Sentiment: Neutral
Legal Complexity of Training AI on Copyrighted Books
Courts and legal experts remain divided over whether training large AI models on copyrighted books constitutes infringement. Recent rulings have produced mixed signals: Judge William Alsup ordered Anthropic to pay a $1.5 billion settlement to writers but ruled the company’s model training itself lawful, penalizing the use of pirated “shadow libraries.” Other cases, such as Thomson Reuters v. Ross Intelligence, found that copying proprietary content to build a competing AI product was not fair use. The 1976 Copyright Act has not been updated for modern AI use-cases, so judges are applying century-old doctrines—especially the fair-use “transformative” test—to novel technical processes. Legal uncertainty persists (including questions about the copyrightability of fully AI-generated works), and ongoing litigation will continue to shape industry practices for model training and content use.
Court rulings and unsettled legal doctrine around training AI on copyrighted works directly affect model development, licensing risk, and content usage practices across the AI and adjacent ad/marketing industries.
Track Anthropic Signals & Market Shifts in Real-Time
Polaris7 autonomous intelligence agents track regulatory filings, primary sources, executive changes, and deal flow 24/7. Create your free Explorer workspace to monitor these entities.
Key Takeaways & Evidence Grounding
- Judge William Alsup ordered Anthropic to pay a $1.5 billion copyright settlement to a group of writers, while finding that Anthropic’s AI training was lawful.
- Alsup penalized Anthropic not for training but for pirating books from illegal online 'shadow libraries.'
- Thomson Reuters sued Ross Intelligence and Judge Stephanos Bibas ruled that Ross’s copying to build a competing AI legal platform was not fair use.
- The Copyright Act of 1976 has not been updated to account for modern AI training practices, forcing courts to apply older doctrines to new technology.
- In Thaler v. Perlmutter, the court ruled that a work that is 100% AI-generated is not copyrightable.
Connected Companies & Entities
5 Entities mapped“Last year, in one of the first rulings of its kind, Judge William Alsup ordered Anthropic to pay a mammoth $1.5 billion copyright settlement...”
“Henderson is referencing a case in which the media and technology company Thomson Reuters sued the research firm Ross Intelligence for copyi...”
“What’s a $1.5 billion fine to a company projecting about [$200 billion] in annual revenue by 2028?...”
““If you write your novel in [Microsoft] Word and run spell check, we kind of feel comfortable with the idea of saying that Word does not own...”
“Cathy Gellis, an attorney with expertise in intellectual property, copyright, and technology, told TechCrunch....”
Ontology Mapping & Concepts
Related Market Signals & Shifts
Recent verified developments and strategic activity across this market segment.
US Government Sides With OpenAI in Copyright Training Lawsuit
The Trump administration has filed a 20-page legal brief in the copyright lawsuit brought by The New York Times against OpenAI, arguing that AI companies should be allowed to train large language models on copyrighted material without explicit permission. The brief asserts that the United States has a strong interest in maintaining global leadership in artificial intelligence and warns against constraining LLM development under what it calls a misunderstanding of fair use doctrine. It is not a ruling; the case is pending in the U.S. District Court for the Southern District of New York. The intervention could influence the ongoing fair-use debate, which centers on whether training AI on copyrighted works is transformative. The article also notes a related case in which Anthropic was ordered to pay a $1.5 billion settlement to writers, but the penalty was tied to its use of illegal shadow libraries, not to the act of training models on copyrighted texts.
Microsoft Managers Call AI Training 'Ridiculous Theft'
In a copyright lawsuit filed by The New York Times against Microsoft and OpenAI, internal documents reveal that senior Microsoft executives privately view AI training on copyrighted content as a massive theft of labor, despite public justifications. Brent Hecht, head of applied sciences at Microsoft, called it "astonishing theft of unprecedented scale," while OpenAI's ChatGPT chief Nick Turley admitted the technology poses an "existential threat" to news publishers. The documents also suggest that internal opinions about the "fair use" defense are skeptical, with concerns that prevailing on it would make the term meaningless. Microsoft executives, including CEO Satya Nadella, acknowledged under oath that chatbots can replace visits to original sources, leading to a potential "vicious circle" harming both models and the internet. The case, still ongoing, is a key precedent for whether AI training qualifies as fair use.
NYT-Opening Documents Reveal AI Scraping, Fair-Use Impact on Publishers
Newly unsealed documents in The New York Times' lawsuit against OpenAI and Microsoft reveal internal admissions that AI products are 'largely substitutive' and pose an 'existential threat' to publishers. Legal experts say this eviscerates the fair-use defense, especially given evidence of paywall circumvention, which may violate the DMCA. The filings could catalyze more publisher lawsuits and force AI companies into paid licensing markets. Executives from OpenAI, Microsoft, and other AI firms are shown acknowledging the commercial value of news content while using it without permission. The case, presided over by Judge Sidney Stein, may set precedent for AI training on copyrighted works, reshaping the terms of trade between AI and publishing.
Track Real-Time Market Signals & Shifts
Set up custom watchlists to receive automated, evidence-grounded executive digests whenever material signals or shifts occur across your tracked landscape.