#1
These AI Companies Used Hundreds Of Songs To Train AI Models; Now Accused Of Stealing Data
Anthropic is once again making the headlines for the wrong reasons and stealing content to train AI models is becoming a worrying trend.
AI companies like Anthropic and Suno have been accused of stealing copyrighted songs and using them to train their AI models which could eventually cost them billions in penalties.
The latest push to penalise AI companies for infringing copyright material comes from a music publisher Round Hill, who claims both Suno and Anthropic are stealing their content an
展开查看正文
Anthropic is once again making the headlines for the wrong reasons and stealing content to train AI models is becoming a worrying trend.
AI companies like Anthropic and Suno have been accused of stealing copyrighted songs and using them to train their AI models which could eventually cost them billions in penalties.
The latest push to penalise AI companies for infringing copyright material comes from a music publisher Round Hill, who claims both Suno and Anthropic are stealing their content and using them to train AI models without seeking permission or authority.
The report by Reuters claims that a case has been filed in a California court where these AI companies will be grilled over their unethical practices to get their models trained and running in the market.
More AI Training With Stolen Data
Round Hill, in its lawsuit against the Claude maker, claims the AI company has taken lyrics from around 500 songs that has helped Claude train better in this subject matter. The publisher says the tracks include popular artists like Goo Goo Dolls, James Brown and The Kinks, enabling the AI chatbot to give answers around prompts that are related to music and these celebrities.
Suno is the other AI company that is mentioned in the lawsuit, and also accused of using their content to train the models.
The publisher is seeking damages for these actions and hopes to receive well over $1 billion from the accused parties with more spotlight on AI companies for their alleged practices.
They have no interest in settling the case and want these billion-dollar valued entities to face penalties for their lack of policies to train AI and use anyone's content without seeking consent or at least letting them know about it.
Anthropic has already been heavy in the news circle lately, not just because of its copyright allegations from other companies.
Even Suno is being challenged by major music brands like Universal Music Group and Sony Music and accused the company of using their content to train the AI systems and make their product grow at the expense of established brands and their library.
At the same time, Anthropic is doing its own exercise of authenticating AI-generated content. The company has detailed a text-watermarking system planned for future Claude models, offering a way to estimate whether the AI assistant was involved in writing or editing a passage.
The watermark will not appear as a visible label or hidden character. Instead, it will create a detectable pattern through the word choices Claude makes while generating text.