YouTube's Automated Copyright Cop Gets Sensitivity Training
Content ID system has been enhanced with an improved appeals process and more accurate content recognition code.
Enterprise Social Networks: A Guided Tour
(click image for larger view and for slideshow)
Google's YouTube has revised its Content ID system to recognize content more accurately, to reduce the incidence of mistakes, and to force copyright holders to either drop a claim or file a formal DMCA notification when a user appeals a rejected Content ID counter-claim.
YouTube rights management product manager Thabet Alfishawi in a blog post said that although the company still has work to do to improve the way Content ID works, these changes will help ensure that "the rights of both content owners and users are protected and everyone can control their original content and make money from it."
YouTube introduced Content ID in October 2007, seven months after Viacom sued Google alleging "massive copyright infringement" on YouTube. Google prevailed in a 2010 ruling and saw that ruling largely upheld earlier this year, but it has nonetheless continued to develop its Content ID technology to catch unauthorized use of copyrighted material on YouTube.
Content ID essentially is a digital fingerprinting system for videos. Content owners provide YouTube with audio and video reference files, along with associated metadata. The owners can then set policies that tell the Content ID system what to do if and when it finds full or partial copies of protected content. Along with allowing content owners to block infringing content, Content ID also provides options for obtaining statistics about viewership of the identified video or making money from the video through the insertion of ads.
Unfortunately, Content ID isn't perfect, and few want their voice on YouTube silenced because of technical errors or the overreach of copyright holders. The Electronic Frontier Foundation has been raising concerns about the technology since its introduction. In 2009, after failed licensing negotiations between Warner Music and YouTube led user-created videos with Warner Music songs to be blocked, the cyber-rights group argued that Content ID "fails to separate the infringements from the arguable fair uses."
The Content ID enhancements introduced by YouTube on Wednesday aim to reduce the likelihood of errors and to ensure that users who contest claims can appeal if the content owner rejects a claim of fair use.
"Content ID abuse is a longstanding problem and it's good to see Google giving its users new tools to challenge improper flags," said EFF intellectual property director Corynne McSherry in an email. "We are still reviewing the changes--and we'll be watching the implementation carefully--but this seems like a step in the right direction."
To reduce the likelihood of an invalid claim, YouTube has improved the algorithms used to identify potentially invalid claims. In addition, human intelligence has been added to the mix: Content screened for invalid claims now gets placed in a queue for a manual review.
The new appeals process gives users an extra appeal. Previously, users were able to dispute Content ID claims, but if the content owner rejected that dispute, the user didn't have any recourse for certain kinds of content, such as videos monetized with ads. Now, users can appeal the rejection of their initial counter-claim. When they do so, the content owner must either release the claim or file a formal DMCA takedown notification.
Although YouTube's improved Content ID system still won't be perfect, perhaps it will slow further additions to the EFF's Takedown Hall of Shame.
InformationWeek Tech Digest, Nov. 10, 2014Just 30% of respondents to our new survey say their companies are very or extremely effective at identifying critical data and analyzing it to make decisions, down from 42% in 2013. What gives?