YouTube has been defending its Content ID system following criticism of its anti-infringement tool in recent submissions to the US Copyright Office by the three major labels.
It represents the latest front in the war of words between YouTube and the music industry at a time when the service is preparing to negotiate new licensing deals – and also when reviews of safe-harbour legislation are underway on both sides of the Atlantic.
The label criticism came in their submissions to the US Copyright Office’s public consultation on the safe-harbour provisions of the US DMCA legislation, first reported on by the Financial Times. Each label claims Content ID is not a foolproof tool for detecting infringement in user-uploaded videos on YouTube.
“Content ID does not do a sufficient job in identifying videos incorporating recordings that are in the Content ID system, and even back in the 2009 timeframe, video creators knew how to defeat detection by Content ID,” claimed WMG’s submission, which also claims that the tool does not identify live performances by its artists.
Universal’s filing includes the claim from its publishing arm that “Content ID fails to identify upwards of 40% of the use of UMPG’s compositions on YouTube”, as well as the note that UMG employs around half a dozen people in the US and UK “solely to address YouTube disputes” when uploaders challenge Content ID claims.
UMG admits that it is paying a third-party firm “hundreds of thousands of dollars per year” in commission to identify infringing YouTube uploads that have not been picked up by Content ID.
Sony’s filing makes a similar claim, with its third-party contractor identifying nearly 1.5m infringements on YouTube missed by Content ID since December 2012: videos that “would have cost Sony and its artists $7.7 million dollars in revenue from approximately 10 billion plays” if undetected.
YouTube is making its defence in public, with figures provided to Music Ally by the company. The company says that more than 98% of copyright management on YouTube now takes place through Content ID – and that for music specifically, 99.5% of sound-recording claims are automated through the system, with only 0.5% claimed manually.
“They set the parameters for their content, and the system manages it for them,” claims YouTube. “The system rarely makes matching mistakes – the system matches audio content with over 99.7% precision – meaning the match we found actually contains the audio we were given,” added the company, which says that less than 1% of Content ID claims are disputed.
The key for labels is whether these tiny percentages – the 0.5% of Content ID cases claimed manually and the 1% of overall claims that are disputed – represent an unfair burden or not.
The bigger picture remains the music industry’s desire for YouTube to be stripped of its safe-harbour protection and what they see as its unfair leverage at the negotiating table for licensing.
But as the arguments above show, this will see every aspect of YouTube’s defence being picked over by labels, while YouTube’s willingness to challenge rightsholders publicly on their own transparency on how its payouts are being distributed to creators is growing.
“Google has paid out billions to the music industry and we’re engaged in productive conversations with the labels and publishers around increasing transparency on payouts. We believe that by providing artists and songwriters greater visibility around revenue earned on YouTube, we can solve many of these issues,” a spokesperson said this weekend.