AI training is theft, court to decide
„AI companies trained on copyrighted music, books, and images without permission or payment, and courts must sanction this with damages.”
Counter-narrative
Training AI is transformative fair use, while mandatory licensing would create a tech oligopoly; this is the position of AI companies and the US Department of Justice.
Carriers
GEMA and collective management organizations, Major labels (UMG, Sony, Warner), Independent music publishers (Round Hill, Reservoir), Authors Guild and writers' organizations, Hollywood agencies and creator coalitions (CAA, WME, writers' guilds), Legal trade press (MLex)
Analysis
The narrative holds that generative models are built on freely acquired creative work, with the price to be decided in court. The pace of lawsuits is dictated by major publishers and independent catalog owners (Round Hill, Reservoir, Eight Mile Style, the rights holders of Elvis Presley). Billion-dollar damage claims serve as both legal and negotiation tools. Unsealed documents, including a Microsoft executive's statement on the "greatest theft of human labor," bolster the rights holders' argument. The weak point is fragmentation. In Munich, GEMA won; in London, Stability AI. In the US, the DOJ argues for fair use, while the appellate court rejected it in the ROSS case. Thus, no unified rule exists. Courts also demand specificity: in the second publisher lawsuit against Anthropic, the judge questioned the specificity of the piracy allegations.
What's at stake: It will be decided whether creative works qualify as paid raw material for training data or as freely usable public goods.
Narrative sub-threads
Music generators in the crosshairs of rights holders
Collective rights management organizations and publishers in several countries are suing Suno and Udio over training materials. Suno admitted to sourcing YouTube audio for training but disputes that UMG and Sony are entitled to enforce 'stream ripping' claims. Udio calls the training 'typical fair use'. A first-instance judgment was issued in Munich against Suno: the court ruled that both the training and the output infringe rights represented by GEMA.
Music-generating startups are at the center here, with the plaintiffs being collective rights management organizations and record labels. The dispute concerns the acquisition of training material (stream ripping) and model output, not song lyrics.
↗ rising · evidence: moderateSong lyrics in Claude's training
Music publishers and catalog owners are suing Anthropic. The allegation is that the company used tens of thousands of songs, including lyrics, without authorization to train Claude. Damages sought are in the hundreds of millions of dollars, with some reports citing billions.
The main target of this branch is a developer of a large language model, though Round Hill and Reservoir are also suing Suno. The charges extend to the use of texts and the removal of author attributions. The judge expressed doubts about the specificity of the piracy allegations and for now rejected the direct infringement lawsuit against Dario Amodei.
▲ at peak · evidence: strongPirated books and news: OpenAI lawsuits
Book authors and news publishers are suing OpenAI, Microsoft, and Meta. According to unsealed court documents, OpenAI trained on books from the pirate collection LibGen. According to the authors' filings, OpenAI and Microsoft executives knew that mass book copying was illegal. Nearly 300 authors are suing Meta; the company says no one 'decided' on piracy. The front is sharpening in the amicus battle over fair use.
Textual works and press content are at the center here. Plaintiffs cite internal corporate documents: LibGen use and a Microsoft executive's statement ("the greatest theft of human labor"). The US Department of Justice argued in support of OpenAI and Microsoft in the New York Times lawsuit.
↗ rising · evidence: strongPrecedents: courts' mixed record
The first court rulings are mixed. The Third Circuit Court of Appeals rejected fair use protection for AI training in the ROSS case. The UK High Court ruled largely in favor of Stability AI in the Getty–Stability dispute. In DMCA-based appellate cases, Microsoft, OpenAI, and GitHub won. The Court of Justice of the EU is examining the issue for the first time; Advocate General Maciej Szpunar stated that Gemini responses may infringe publishers' rights.
This branch is not about new lawsuits but about decisions that set the direction of legal application. The actors are courts and advocates general; the evidence is the judgment or advocate general's opinion, not training data.
→ lasting · evidence: moderateMusicians against their own labels
The American Federation of Musicians (AFM) is suing its own major labels. It claims they handed over recordings for AI training without consent or compensation. Universal and Warner seek dismissal of the lawsuit, while the AFM demands AI protection in contract negotiations under the slogan "Human Made, Human Paid".
Here, creators face not AI companies but publishers that license their rights to AI firms. The debate concerns how much musicians can influence these agreements and whether they receive compensation.
Sub-thread stories
Music generators in the crosshairs of rights holders
Song lyrics in Claude's training
Pirated books and news: OpenAI lawsuits
Precedents: courts' mixed record
Musicians against their own labels
Google search interest
Hungary
| Topic and search term | Weekly | Average |
|---|---|---|
| AI training is theft, court to decideartificial intelligence copyright | 0 | |
| Music generators in the crosshairs of rights holderssuno ai zene | 0 | |
| Song lyrics in Claude's traininganthropic per | 0 | |
| Pirated books and news: OpenAI lawsuitsopenai per | 0 | |
| Precedents: courts' mixed recordfair use | 0 | |
| Musicians against their own labelsmusicians' union | 0 |
World
| Topic and search term | Weekly | Average |
|---|---|---|
| AI training is theft, court to decideai copyright lawsuit | 3.5 | |
| Music generators in the crosshairs of rights holderssuno lawsuit | 2.5 | |
| Song lyrics in Claude's traininganthropic music lawsuit | 0.2 | |
| Pirated books and news: OpenAI lawsuitsopenai copyright lawsuit | 0.1 | |
| Precedents: courts' mixed recordfair use ai | 12.4 | |
| Musicians against their own labelsmusicians union ai | 0.1 |
Google Trends relative search index over the period studied, weekly average. A narrative and its sub-threads appear in a single comparison: 100 is the group's most-searched day, and the figures can only be compared within the group. The search term was chosen by machine analysis; it measures how much the term is searched, not how widespread the narrative is.
Hungary
The narrative has little or no domestic footprint during the period examined: there is no Hungary-related item among the stories.
Weekly trend
Turning points
Related narratives in the News magazine
Supporting stories
Machine analysis based on story titles and summaries, with a verification pass. The source of the facts is the article linked at each story.