Anthropology has An agreement has been reached. and agreed to stop providing users with lyrics based on copyrighted songs from various music publishers. Back in 2023, AI companies were under at Universal Music Group, Concord Music Group and others after discovering that Claude's chatbot would return lyrics to songs like Beyoncé's “Halo” when prompted.
The entertainment industry is one of the most litigious industries and fights vigorously to protect its copyrights. Let's go back and look at historical cases. From the destruction of Napster to Viacom's years-long legal battle with YouTube, most recently the popular lyric annotation site Rap Genius (now called Genius). under by the National Music Publishers Association to produce copyrighted song lyrics.
The music publisher suing Anthropic admitted that other websites, such as the Genius song annotation platform, published lyrics online. But it noted that Genius eventually began paying licensing fees to publish on its website.
In this latest case The music publisher claims that Anthropic copied the lyrics from the web and intentionally removed the watermark placed on the lyrics website. To help identify where copyrighted content was published, after Genius began requesting lyrics copyright from music publishers. It was also a smart thing. Insert additional apostrophes into the lyrics so that in the event that content is improperly copied, Genius will know that clearly paid for content has been stolen and can demand its removal.
Anthropic does not accept such claims. But as part of the deal, it agreed to better maintain guardrails to prevent AI models from infringing on copyrighted material. It will also work in good faith with music publishers when it discovers that guardrails are not working.
anthropology protected Using song lyrics and other copyrighted content to train AI models counts. hollywood reporter“Our decision to make this provision is consistent with those priorities. We still aim to show that. To comply with existing copyright laws The use of potentially copyrighted content to train general AI models constitutes substantive fair use.” This argument is at the heart of AI companies' defense of copyrighted content appearing in their models. Supporters claim that remixing copyrighted content from websites such as new york times Fair use is considered fair use as long as a significant change is made through derivative works.
News and music publishers disagree. And the lawsuits against Anthropic don't end there. The music publisher is still seeking a court order to prevent Anthropic from training future models on any copyrighted lyrics.
Concerns about infringement arose from the possibility that Anthropic's models would be used in the creation of music. This causes musicians to lose control of their art. It's not an unfounded concern. This is because there was widespread speculation that OpenAI was imitating Scarlett Johansson's voice after she refused to provide her voice for its AI voice model.
Tech companies like OpenAI and Google monetize their platforms and network effects. Not from selling copyrighted content. This has always led to tension between Hollywood and Silicon Valley. Art is just “content” meant to serve the larger purpose of generating engagement and selling ads. The AI slop that fills Facebook today is representative of how tech companies view everything that can be done. Can be used interchangeably
Publisher likes time Perplexity has fought high-profile battles with OpenAI in court to stop them from owning copyrighted content. OpenAI has tried to respond by seeking licenses from some companies and other AI players. Perplexity has begun testing a revenue sharing model. can But publishers want more control and not be forced into these shaky deals that could end at any time and still drive people off their sites. In other words, this is far from over. From the end of the story when it comes to disputes over copyrighted material in large language models.