发生了什么
Billboard reports that Sony Music Publishing and Warner Chappell Music filed a copyright lawsuit against Anthropic, its CEO Dario Amodei and co-founder Benjamin Mann. The publishers allege that Anthropic used more than 20,000 songs to train its AI models and acquired some works through torrenting, scraping and downloading. Anthropic disputed the claims and said it would defend the case.
Billboard reports that Sony Music Publishing and Warner Chappell Music filed a new copyright complaint against Anthropic on Aug. 28. The lawsuit also names Anthropic CEO and co-founder Dario Amodei and co-founder Benjamin Mann as defendants. According to Billboard, the publishers allege that Anthropic “illegally harvested” more than 20,000 musical compositions to train its artificial-intelligence models, including works such as “Ain’t No Mountain High Enough,” Mariah Carey’s “All I Want for Christmas is You” and Taylor Swift’s “Paper Rings.”
Billboard reports that the publishers’ complaint alleges Anthropic acquired copyrighted music through “torrenting, scraping, and downloading” on a large scale. That allegation concerns not only the use of the songs in model training but also the way the material was obtained. The publishers characterize Claude as having been built through copyright infringement, while also saying they recognize the potential of ethical AI technology and arguing that such systems must operate within legal limits.
Anthropic told Billboard that the new case is the third lawsuit from the same lawyers and that it recycles allegations already before the courts. The company said training models is transformative fair use and that it would defend itself robustly. Billboard reports that Anthropic’s position conflicts with the publishers’ claim that the company’s conduct amounts to large-scale infringement.
The lawsuit follows earlier music-related litigation against Anthropic, including cases brought by Universal Music Group, Concord, BMG and Round Hill Music, according to Billboard. The report also places the dispute within a broader wave of lawsuits against AI companies over training data. Billboard says the source of the songs, the legality of training, and possible damages remain central issues, but it does not provide the complaint’s docket number, a court ruling on this new case, or independent verification of the allegations.
为什么这很重要
The case adds two major music publishers to a growing legal fight over whether copyrighted works can be used to train models. Billboard reports that the central fair-use question remains unresolved, while a prior ruling found that using pirated materials can create separate liability. The outcome could affect licensing, training practices and potential damages across the AI industry.
The dispute goes to a foundational question for : whether a company may copy copyrighted works into training systems without permission when the resulting model is intended to generate new material. Billboard reports that AI companies argue this is a legal form of transformative fair use that supports innovation, while copyright owners argue that mass copying and commercial use require authorization. The report says the fair-use issue is still largely undecided, leaving billions of dollars in damages and licensing value at stake.
The case also highlights a distinction between training on copyrighted material and obtaining that material unlawfully. Billboard reports that a judge ruled in an earlier case that even if AI training itself is lawful, using pirated materials is not. The publishers cite that ruling in their new complaint and argue that Anthropic’s alleged acquisition methods provide an independent basis for liability. The source does not independently confirm the prior ruling’s full scope or how it would apply to this case.
For music publishers and songwriters, the practical stakes include control over catalogs, compensation for use in AI development and the possibility that generative systems could compete with or alter existing creative markets. Those consequences are potential implications of the litigation, not established outcomes. Billboard reports that the publishers say the alleged conduct caused immense harm, but the source does not quantify losses to Sony, Warner or individual songwriters.
For AI companies, the case may influence whether training data is licensed, filtered, documented or obtained through unauthorized channels. A ruling or settlement could shape commercial negotiations between model developers and rights holders, but Billboard does not report that this lawsuit has changed Anthropic’s products, access terms or training practices. It also does not establish that Claude generated unauthorized copies of the named songs or that users were exposed to them.
互动机制：它实际上是如何运作的
以交互方式探索这一发展背后的基础技术。
crm_get_transaction(id='4092').Which component of an AI application is the machine-learning model itself?
接下来看什么
Watch for Anthropic’s formal court filings, any response from Sony or Warner, and developments in the underlying fair-use and piracy questions. Billboard does not independently confirm the publishers’ allegations, the precise evidence supporting them, or the status of any earlier settlement beyond what it reports. The source also does not establish whether Claude outputs reproduced protected musical expression.
The next important developments are likely to come through Anthropic’s court response and any judicial decisions on the complaint. Those filings could clarify which songs are at issue, what evidence the publishers have regarding acquisition and training, and whether the defendants challenge the case on procedural or substantive grounds. Billboard’s report does not say when Anthropic must respond or whether the court has scheduled a hearing.
Watch whether the parties pursue licensing discussions, settlement negotiations or additional claims. Billboard reports that Anthropic paid a $1.5 billion settlement in an earlier case after a piracy ruling, and that Sony and Warner cite that outcome as evidence that further legal action is needed. The source does not independently verify the settlement terms, identify the earlier case by docket number or explain whether the new complaint seeks a specified amount of damages.
Another issue is how courts distinguish between a model’s training process and its outputs. Billboard describes the industrywide fair-use dispute but does not report evidence that Claude reproduced the specific compositions named in the complaint. Future proceedings may address whether copying during training, storing unauthorized files, generating similar material or distributing outputs create separate legal consequences.
Meaningful unknowns remain. Billboard does not independently confirm the allegation that more than 20,000 works were used, the alleged torrenting and scraping campaign, or the publishers’ characterization of Claude’s development. The report also does not include a response from Sony or Warner beyond their complaint, any position from songwriters, or a ruling on the merits. Those facts will matter before the allegations can be treated as established findings.