Posted by: Rob Hof on October 15, 2007
YouTube, under fire from movie studios and television networks for letting pirated versions of their creations loose on the popular video site, finally is announcing a system to help content owners prevent piracy in a somewhat more automated way than today, when they have to hire people to find them manually. But whether they’ll go along with the Google-owned video site’s plan remains uncertain.
The YouTube Video Identification system, just announced on the official Google blog, basically will require studios to upload any videos they want to claim as their own. Then YouTube, using Google video I.D. technology under development for several years, will create an I.D. file, so that when someone uploads a video, it can be matched and removed. Although YouTube says that for now, the system may not be able to prevent uploading, files won’t be indexed until they’re run through the filtering process. Eventually, YouTube hopes to be able to block pirated uploads before they post at all.
That is, if the studios want them blocked. When they provide a video to YouTube, the studios can specify whether it should be blocked, promoted as is, or "monetized" with ads or other moneymaking schemes YouTube may come up with.
One of the challenges here is that the system requires studios and others to provide YouTube/Google with all the videos they want to have blocked, promoted, or made available with advertising attached. That could be a gargantuan task in itself, at least at first. And given that YouTube has antagonized some studios by not filtering pirated clips itself, they may not trust Google with their crown jewels. YouTube says it won't use the videos for anything except creating I.D. files, and it says studios can choose to use the same tools Google developed for YouTube to create the I.D. file themselves and submit that to YouTube, though that's a longer process.
YouTube claims it can't know in advance which videos are intended to be uploaded by studios and which it wants blocked unless the studios tell it specifically. "We really need the content community to work with us," said David King, a YouTube product manager. "We need something to match (the video at issue) to. We can't automatically match in thin air."
The system is going out in beta, and YouTube wouldn't provide data on how well it works so far. During a brief demo, officials said that 18 matches were made in tests in the past week and a half. That's right, 18, not 18,000.
Since I was just briefed on the system, I haven't had a chance to check in with Time-Warner and Disney, two of the nine companies Google and YouTube have talked to about the system. YouTube claims response has been "very positive," but we'll have more on that later as I and my colleague Peter Burrows dig in further. It's hard to tell just yet whether this is a move that Google and YouTube really expect big studios to go along with, or a maneuver to help defend against the $1 billion infringement lawsuit filed by Viacom last year.
Meanwhile, Google isn't alone in the video I.D. area, with startups such as Vobile angling to become the standard. And trust issues, or at least legal entanglements, could make it tough for Google to gain the studios' favor. Asked if Google had discussed the system with Viacom, YouTube Chief Counsel Zahavah Levine replied, "Um," paused for a few seconds, and said only that the system will be available to Viacom as well as every other content creator.