Under the proposed law, this “content provenance information” would be embedded in digital forms of copyrighted material, and it would be unlawful to remove it or tamper with it, except in very limited cases where platforms are carrying out research to improve security. It would also be unlawful for anyone to use any material with “content provenance information” to train AI, or to create AI-generated content, “unless such person obtains the express, informed consent of the person who owns the covered content.”
Rights
Apple, Nvidia, Anthropic Used Thousands of Swiped YouTube Videos to Train AI
AI companies are generally secretive about their sources of training data, but an investigation by Proof News found some of the wealthiest AI companies in the world have used material from thousands of YouTube videos to train AI. Companies did so despite YouTube’s rules against harvesting materials from the platform without permission.
Source: Apple, Nvidia, Anthropic Used Thousands of Swiped YouTube Videos to Train AI
CCC Launches Collective Licensing for AI
Copyright Clearance Center (CCC) has launched a collective licensing solution for the internal use of copyrighted materials in AI systems. The new service, which became available July 1, is an addition to CCC’s existing Annual Copyright Licenses (ACL) service and aims to provide a streamlined method for companies to access a consistent set of rights across multiple rightsholders, while ensuring compensation for content creators.
Licensing Your Movie & TV Content for AI Training: Can You? Should You?
As the most obvious advantage for library owners, granting training data licenses creates an additional revenue stream for existing programming. However, without suitable implementation, the potential cons for the licensing library owner include harming relationships with creative collaborators and partners whose persona, work or assets are ingested into the AI model along with the licensed library programming.
Source: Licensing Your Movie & TV Content for AI Training: Can You? Should You?
An Estimated 73% of YouTube Music Subs Use Bundled Plans
Spotify’s main competitors – it along with Apple Music, Amazon Music, and YouTube Music collectively account for the vast majority of music streaming subscribers in the States – haven’t been as brazen in adopting bundles. But as demonstrated by publisher royalty statements shared with and vetted by DMN, they also rely heavily on multi-product packages. For YouTube Music, the fourth-largest on-demand music platform in the U.S. by paid users, that refers to the estimated 72.99% of subscribers who access the service via a bundled plan.
Source: An Estimated 73% of YouTube Music Subs Use Bundled Plans
PPL: ‘Neighboring rights are becoming a significant and vital revenue stream’
Sound recording performance rights, also known as neighboring rights income, reached $2.7 billion in 2023, representing 9.5% of the global recorded music market. While ‘Old Europe’ has traditionally been the hub of these types of revenue streams – namely, rights from TV and radio broadcasting; aligned digital uses such as simulcasting, webcasting and satellite radio; and sound recordings played in public venues – a significant shift is underway.
Lucian Grainge Isn’t Mincing Words on AI Music and IP Theft
Grainge says the advent of AI technology has reached a point where Universal Music must “be completely at the epicenter of its application. An example he uses is the Beatles’ 2023 single “Now and Then” which used AI to isolate and clean up an old recording of John Lennon singing. “It’s a brilliant song—great lyrics, fabulous performance, incredibly emotive—that unless we’d had AI to individualize different recordings, would have never come to light, “Grainge told the L.A. Times.
Source: Lucian Grainge Isn’t Mincing Words on AI Music and IP Theft
Beatport and Beatdapp team up to tackle music-streaming fraud
Anti-fraud startup Beatdapp continues to make headlines with its data on just how many music streams might be illegitimate – most recently in a Sky News story suggesting that criminals might be making up to $3bn a year from streaming fraud. Now the company has announced its latest partnership with a music service trying to tackle this. Electronic music-focused DSP Beatport is going to be using Beatdapp’s fraud-detection technology.
Source: Beatport and Beatdapp team up to tackle music-streaming fraud
YouTube will use AI to snip copyrighted music and not silence your whole video
YouTube is turning to artificial intelligence to try to simultaneously appease copyright holders of songs while making life a little easier for those who upload videos with songs they don’t have permission to use. Instead of just taking down a video with copyrighted audio the uploader doesn’t own, they can use a new AI tool to remove the protected song without erasing the rest of the video’s audio track.
Source: YouTube will use AI to snip copyrighted music and not silence your whole video
Warner Music Warns AI Companies About Unlicensed Training
Warner Music Group has formally warned AI companies of training their models on protected media and other assets without a license. WMG forwarded the relevant notice to a number of high-profile artificial intelligence players, several of which are grappling with litigation over alleged copyright infringement. And that litigation centers mainly on the media, apparently including copyrighted works, used to train the underlying AI systems.
Source: Warner Music Warns AI Companies About Unlicensed Training