US judge: AI training on books is fair use, partly
SAN FRANCISCO (United States): A US federal judge has sided with Anthropic regarding training its artificial intelligence models on copyrighted books without authors' permission, a decision with the potential to set a major legal precedent in AI deployment.
District Court Judge William Alsup ruled on Monday that the company's training of its Claude AI models with books bought or pirated was allowed under the 'fair use' doctrine in the US Copyright Act.
'Use of the books at issue to train Claude and its precursors was exceedingly transformative and was a fair use,' Alsup wrote in his decision.
'The technology at issue was among the most transformative many of us will see in our lifetimes,' Alsup added in his 32-page decision, comparing AI training to how humans learn by reading books.
Tremendous amounts of data are needed to train large language models powering generative AI.
Musicians, book authors, visual artists and news publications have sued various AI companies that used their data without permission or payment.
AI companies generally defend their practices by claiming fair use, arguing that training AI on large datasets fundamentally transforms the original content and is necessary for innovation.
'We are pleased that the court recognized that using 'works to train LLMs was transformative,'' an Anthropic spokesperson said in response to an AFP query.
The judge's decision is 'consistent with copyright's purpose in enabling creativity and fostering scientific progress,' the spokesperson added.
- Blanket protection rejected -
The ruling stems from a class-action lawsuit filed by authors Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson, who accused Anthropic of illegally copying their books to train Claude, the company's AI chatbot that rivals ChatGPT.
However, Alsup rejected Anthropic's bid for blanket protection, ruling that the company's practice of downloading millions of pirated books to build a permanent digital library was not justified by fair use protections.
Along with downloading books from websites offering pirated works, Anthropic bought copyrighted books, scanned the pages and stored them in digital formats, according to court documents.
Anthropic's aim was to amass a library of 'all the books in the world' for training AI models on content as deemed fit, the judge said in his ruling.
While training AI models on the pirated content posed no legal violation, downloading pirated copies to build a general-purpose library constituted copyright infringement, the judge ruled, regardless of eventual training use.
The case will now proceed to trial on damages related to the pirated library copies, with potential penalties including financial damages.
Anthropic said it disagreed with going to trial on this part of the decision and was evaluating its legal options.
'Judge Alsup's decision is a mixed bag,' said Keith Kupferschmid, chief executive of US nonprofit Copyright Alliance.
'In some instances AI companies should be happy with the decision and in other instances copyright owners should be happy.'
Valued at $61.5 billion and heavily backed by Amazon, Anthropic was founded in 2021 by former OpenAI executives.
The company, known for its Claude chatbot and AI models, positions itself as focused on AI safety and responsible development.
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
The Sun
3 hours ago
- The Sun
Vietnam tycoon will not face death penalty over $27 bn fraud: lawyer
HANOI: A Vietnamese property tycoon sentenced to death for fraud totalling $27 billion will no longer face execution, her lawyer said Wednesday, as the country abolished capital punishment for eight crimes, including property embezzlement. Property developer Truong My Lan's lawyer Giang Hong Thanh told AFP that her death penalty 'will be converted to life imprisonment... I informed Ms. Lan this morning, she is very happy.' Vietnam on Wednesday abolished the death penalty for eight crimes including espionage, graft and attempting to overthrow the government, according to state media. Lan, 68, was convicted last year of swindling money from Saigon Commercial Bank (SCB) -- which prosecutors said she controlled -- and sentenced to death for fraud with damages totalling $27 billion -- equivalent to around six percent of the country's GDP. She appealed the verdict in a month-long trial, but in December the court in Ho Chi Minh City determined that there was 'no basis' to reduce her sentence. But the court said Lan's sentence could be reduced to life imprisonment if she returned three quarters of the stolen assets. 'If Ms Lan compensates for three quarters of the losses along with some other conditions, she will continue to be considered for a further reduction of her sentence,' lawyer Thanh said Wednesday. Lan 'is still actively cooperating with state agencies of Vietnam and her partners to find ways to deal with (her) assets and restructure SCB to come to a final solution to the case.' Lan owned just five percent of shares in SCB on paper but courts concluded that she effectively controlled more than 90 percent through family, friends and staff. Tens of thousands of people who had invested their savings in the bank lost money, prompting rare protests in the communist nation. In a separate $17 billion money-laundering case, Lan had a life imprisonment sentence cut to 30 years on appeal in April. According to the amended penal code approved by the National Assembly Wednesday, the eight crimes no longer punishable by death include property embezzlement and bribe-taking. Under Vietnamese law, those sentenced to death for the eight crimes -- which also include spying and anti-state activities -- before July 1 will have their sentences converted to life imprisonment by the chief judge of the Suprem People's Court. Minister of Public Security Luong Tam Quang said 'the current structure of capital punishment was problematic and, in some cases, misaligned with evolving socio-economic conditions and the realities of crime prevention.' One of the reasons for the move given by Minister of Justice Nguyen Hai Ninh was that in most cases people sentenced to death for the above crimes were not actually executed. The number of executions has not been made public, but Amnesty International estimates more than 1,200 people were on death row in Vietnam by the end of 2024.
The Star
3 hours ago
- The Star
Japan to step up fight against counterfeit anime goods overseas
The Japanese government will join hands with foreign authorities to investigate and identify distributors of such fake goods and take legal action. - AFP TOKYO: Japan is planning to deepen cooperation with foreign authorities to crack down on counterfeit anime goods, in a bid to better protect intellectual property. To combat the sale of counterfeit goods, the government will cooperate with the Content Overseas Distribution Association, an organisation that tackles the issue of pirated publications. With China in mind, the government will join hands with foreign authorities to investigate and identify distributors of such fake goods and take legal action. The government's policy is part of its Entertainment and Creative Industries National Strategy, which was formulated for the first time to promote the Japanese content industry. The industry includes fields such as anime and video games. The policy aims to support the global expansion of the Japanese content industry as it expects the global content market to grow. The government has set a goal of increasing annual overseas sales of the Japanese content industry to ¥20 trillion (US$138 billion) by 2033, more than three times the current level. It has outlined 100 measures covering 10 priority areas in the strategy, which was revealed by the Economy, Trade and Industry Ministry on June 24. The overseas advancement of video games will be supported by the government through the creation of a database that lists the legal regulations, cultural norms and customs of various countries. For manga, the government will support the development of translation tools utilising generative artificial intelligence. The government will also work on the revitalisation of bookstores in Japan, whose numbers have been decreasing rapidly in recent years, as they serve as important sales base of manga. It also considers support to manga to be important on the grounds that many manga titles receive anime adaptations. To address the shortage of anime creators, the government will improve working conditions at production studios. It will also establish a certification mark that will be awarded to anime produced under appropriate working conditions. - The Japan News/ANN
The Star
3 hours ago
- The Star
Bosch CEO warns Europe against regulating 'itself to death' on AI
AI (Artificial Intelligence) letters and robot hand are placed on computer motherboard in this illustration taken, June 23, 2023. REUTERS/Dado Ruvic/Illustration/File Photo FRANKFURT (Reuters) -Bosch CEO Stefan Hartung warned on Wednesday that Europe risks hindering its progress in artificial intelligence compared with other parts of the world through over-regulation. "Europe is unnecessarily delaying its AI future with excessive regulation," said Hartung at a tech conference the German firm is hosting in the south-western city of Stuttgart. Bosch, which holds most of the AI patents in Europe, plans to invest an additional 2.5 billion euros ($2.90 billion) by the end of 2027 in the field. The German car parts supplier is working on AI solutions for autonomous driving and more efficient industry systems, among other applications. The combination of bureaucracy and strict, but also vague, requirements makes Europe comparatively less attractive, said the CEO. Hartung recommended that lawmakers limit themselves to a framework covering only the most important points. Otherwise "we will regulate ourselves to death, because we are trying to regulate against technological progress," he said. Pressure to keep up in the race to develop AI has grown since U.S. President Donald Trump announced a private sector investment of up to $500 billion to fund infrastructure for AI in January. The EU shortly after unveiled plans to mobilise up to 200 billion euros for the sector. ($1 = 0.8618 euros) (Reporting by Ilona Wissenbach and Hakan ErsenWriting by Miranda MurrayEditing by Rachel More)
