🔥 Trade with Pros on Discord → 21 Days Free (No Card)JOIN FREE

Anthropic under fire for alleged AI-fabricated citation in copyright lawsuit

In this post:

  • Music companies claim Anthropic used AI to create a fake academic citation in court.
  • A judge has ordered Anthropic to respond to the allegation, which it calls an unintentional error.
  • This case is part of a larger wave of copyright lawsuits against AI firms over unauthorized use of creative content.

A federal judge in San Jose, California, recently ordered Anthropic, an artificial intelligence company, to address allegations that, in defending against copyright claims by a group of music publishers, it submitted a court document that included an AI-generated “hallucination.”

The lawsuit filed by the music publishers is one of many high-profile cases in which copyright holders are suing tech companies for allegedly using their creations improperly to train AI systems.

Music companies sue Anthropic for alleged AI-fabricated copyright citation

During a hearing in a lawsuit accusing Anthropic of improperly using song lyrics to train its Claude chatbot, a lawyer representing Universal Music Group, Concord, and ABKCO told U.S. Magistrate Judge Susan Van Keulen that an Anthropic data scientist cited a fabricated academic article to support the company’s position in an evidence dispute.

Matt Oppenheim, the music companies’ lawyer, said he had verified the article’s nonexistence with the journal and one of the supposed authors. He referred to the citation as a total fabrication.

Oppenheim said he did not believe the expert, Olivia Chen, intentionally fabricated the citation but noted that they believe that Ms. Chen likely used Anthropic’s AI tool, Claude, to generate her argument and the supporting authority.

As a result of the allegations, Van Keulen requested a response from Anthropic by Thursday regarding the charge, which the company claimed was an unintentional citation error. 

See also  UK lawmakers push for AI stress tests to avoid market disruptions

Interestingly, the music companies asked the judge to question the expert immediately. Still, she disputed the claims, stating that the allegation was “a serious and grave matter” and that there was a significant difference between an AI-generated hallucination and a missed citation.

On the other hand, Sy Damle, Vice Chair of Artificial Intelligence Practice at Latham & Watkins and former General Counsel at the US Copyright Office, expressed dissatisfaction during the hearing, claiming that the accusers were “sandbagging” them by not making the attack sooner. He claimed that although the citation was inaccurate, it referred to the right article.

To support specific criteria for assessing how often Claude reproduced copyrighted song lyrics—a phenomenon Anthropic describes as a ‘rare event’—the expert’s filing cited an article from the journal American Statistician. However, the provided link directed to a completely different article from the same journal, with a different title and authors.

After the hearing, Anthropic’s lawyers and representatives did not immediately reply to a request for comment. Chen also could not be reached for comment right away.

Ziff Davis joins the growing trend of high-stakes copyright lawsuits 

Ziff Davis, a digital media publisher, filed a lawsuit against OpenAI in a federal court in Delaware last month, alleging that the Microsoft-backed artificial intelligence firm had improperly used its publications to train the model for the well-known chatbot ChatGPT.

See also  Can Singapore Emerge as the Premier Global AI Hub? Perspectives from a Google Exec

According to a copy of the lawsuit that the media company provided, Ziff Davis contended that OpenAI deliberately and persistently exploited copyrighted content for its AI systems.

In a statement, the complaint stated, “OpenAI seeks to move fast and break things on the assumption that the federal courts will not be able to effectively redress content owners’ sometimes existential concerns before it is too late.” 

Ziff Davis’s publications include the advice website Lifehacker and the tech news sites ZDNet, PCMag, CNET, and IGN. Other news outlets like Dow Jones and the New York Times have filed copyright theft lawsuits against AI companies.

According to OpenAI and other defendants, such as Google and Meta Platforms, their AI systems legitimately utilize copyrighted content by learning from it to produce original, revolutionary content.

News organizations, writers, visual artists, and others have filed high-stakes copyright lawsuits against OpenAI and other tech companies for allegedly using thousands of copyrighted works to train their generative AI systems without authorization. This new lawsuit joins the growing trend of these cases.

Sharpen your strategy with mentorship + daily ideas - 30 days free access to our trading program

Share link:

Disclaimer. The information provided is not trading advice. Cryptopolitan.com holds no liability for any investments made based on the information provided on this page. We strongly recommend independent research and/or consultation with a qualified professional before making any investment decisions.

Most read

Loading Most Read articles...

Stay on top of crypto news, get daily updates in your inbox

Editor's choice

Loading Editor's Choice articles...

- The Crypto newsletter that keeps you ahead -

Markets move fast.

We move faster.

Subscribe to Cryptopolitan Daily and get timely, sharp, and relevant crypto insights straight to your inbox.

Join now and
never miss a move.

Get in. Get the facts.
Get ahead.

Subscribe to CryptoPolitan