Claude simulated a legal quote. A lawyer had to clean it.

A lawyer defending Anthropic had to clean after the company’s bot ai, appealing to “an embarrassing and involuntary error”.

In a copyright trial on the use of the words of music by Anthropic, the company’s legal team used its AI assistant, Claude, to help write a quote in an expert report.

Claude provided the title of publication, the year of publication and the correct link with the source provided, but “an inaccurate title and incorrect authors,” the lawyer of Anthropic said on Thursday in a judicial file.

Lawyer Ivana Dukanovic, from Latham & Watkins, said that the “manual quotation check” had failed to catch the error and “additional wording errors introduced in quotes during the formatting process using Claude.ai”.

“It was an honest quotation error and not a manufacture of authority,” wrote Dukanovic.

Music publishers Universal Music Group, Concord, and Abkco continued Anthropic, saying that the company used words protected by copyright to train Claude. The case is part of a wave of legal battles between copyright holders and IA companies.

The publishers’ lawyer told court on Tuesday that the anthropogenic data scientist Olivia Chen may have used a false source generated by AI to support the company’s argument, Reuters reported.

Thursday, Dukanovic replied that Chen cited a real article from the journal “The American Statistician”, but Claude had composed the title and the authors.

Anthropic and Dukanovic did not respond to a request for comments.

AI in the legal world

This is not the first time that AI tool has been getting eyebrows in the legal world.

In March, a man without a lawyer deployed an avatar generated by the AI ​​to argue his civil appeal case in a New York courtroom. A panel of amazed judges quickly killed him.

AI hallucinations also landed with hot water lawyers. A lawyer was dismissed from the Baker Law Group after using Chatgpt to generate legal quotes, which turned out to be false.

The former fixer of Donald Trump, Michael Cohen, also was in trouble when he used the Chatbot AI of Google, Bard, to find legal affairs to support his arguments. The chatbot invented the business and his lawyer deposited them in court without checking.

Daniel Shin, deputy research director at the Center for Legal and Court Technology at Virginia’s William & Mary Law School, told Business Insider in a report last month that judges were concerned about the use of AI in courts due to hallucinations.

The courts have shown that they will not tolerate any inappropriate use of AI tools, said Shin.

However, lawyers are told that they must start to adopt AI.

During a conference in legal technology in March, lawyers were invited to adopt AI or risk late, Melia Russell de Bi reported.

“Lawyers must wake up,” said Todd Itami, a lawyer for the major legal defense company Covington & Burling, adding that learning to use artificial intelligence was “imperative” for their success.