|
Internet search giant Google is working to stop the proliferation of illegal online video content. The company's YouTube Web site,  which it purchased last year in a stock deal worth 1.65 billion dollars, has been rife with illegally posted videos taken from commercial media companies since even before Google's purchase. On Monday Google officially released an ambitious and much-anticipated tool for combating illegally posted video content, an endeavor in the beta testing stage called YouTube Video Identification. How Video Identification Works The system aims to identify complex identification signatures unique to every segment of video even before a YouTube user is able to add a video to the Web site, and goes well beyond the company's current system in which illegally uploaded material can stay online until it is reported to YouTube. Google is offering the new YouTube Video Identification system to all media content providers, from small documentary makers to commercial television and film companies. The system will attempt to identify copyrighted material based on key unique characteristics in original digital files provided beforehand from media content providers, and will allow each content provider to decide how an identified video should be handled by YouTube. Choices available to the owners of video content include: - Complete exclusion from YouTube
- Allowing the content to reside on YouTube for promotional purposes
- Licensing the content for use on YouTube and entering into an agreement to share in revenue gained though advertising based on that content
The video identification technology announced yesterday has been developed by Google's own engineers, and a key part of the  system is a requirement that media companies provide digital copies of their content to Google for analyzing and creating the unique identifying digital signatures needed in order for the system to work. While there has been intense debate over whether the burden of enforcing video copyright infringement lies with media companies or with Internet video sharing services, Google's new system seems to be signaling a new age in which media companies may choose to automatically send all their video content to Google in an effort to beat the public to the uploading punch. Once a segment of video content is analyzed by the new Google program, its unique identification signal is stored in a database, against which all future uploads are compared. Should a video being uploaded to YouTube match a known identification signal, the upload would either not be permitted for inclusion on YouTube or would be allowed for posting, depending on the previously defined wishes of the media company or person to whom the work's copyright belongs. The system also attempts to identify video material that has been purposely modified by those looking to thwart such filtering systems, and to identify known material which has been captured by such means as videotaping a television broadcast. In the cat-and-mouse game played by video pirates and the legal owners of video content, it is not difficult to imagine that as soon as the current system in place at YouTube is understood, utilities for adding just enough video interference or extraneous material to trick the system can be far behind. Pending Litigation against YouTube Media companies have become increasingly upset with the increasing availability of copyrighted video content on Web sites such as YouTube, and in March entertainment giant Viacom filed a copyright infringement lawsuit against YouTube and Google for $1 billion. Google is also the defendant in a class-action copyright violation suit filed by the Football Association League of England. Google sees the new video identification program as an additional measure in its efforts to combat illegally uploaded video content, coming on top of a number of programs and policies the company has long had in place for this purpose. Since its inception YouTube has allowed content owners to remove unauthorized video clips, and Google has a strict "repeat-infringer" policy that can terminate those who repeatedly violate the rules. According to the official Google blog, the newly announced YouTube program "goes above and beyond our legal responsibilities." What Media Companies Are Saying Media companies appear willing to give YouTube a chance to try the new video identification system, which they will be monitoring closely as has been reported by New York City-based Viacom general counsel Michael Frickas in a recent New York Times article. "We’ll be watching to ensure that the system is reasonably effective and sufficiently robust to address the issue,” said Frickas, who also expressed some optimism regarding the YouTube Video Identification program. "We are delighted that Google appears to be stepping up to its responsibility and ending the practice of profiting from infringement,” Frickas said. It is not clear whether, should the new program prove successful, Viacom's pending litigation against YouTube and Google might be dropped. What YouTube is Saying The program announced yesterday has undergone testing by nine large media companies¹, including: - NBC Universal
- Disney
- Time Warner
- Viacom
- CBS
YouTube hopes to continue working with what it calls the "content community," as company Product Manager David King noted recently. "We really need the content community to work with us. What really drives this whole thing is having access to the reference material,” said King. YouTube has stated three main goals for the new video identification program according to their announcement, those being: - Accurate information
- Choice for copyright holders
- A great user experience
YouTube claims that early tests with some of the companies that provide video content have "shown very promising results," although no specific figures have been released. An Inherently Imperfect Technology In an interview with several reporters last week, Google CEO Eric E. Schmidt noted that video identification technology may never achieve 100 percent perfect results, but said that even a less than perfect system could still be very useful. "The question is, can we get to 80 or 90 percent?” said Schmidt. While Google has not disclosed exactly how effective the new system currently is at identifying known copyrighted video material, it calls the system "brand-new, cutting-edge stuff," on the YouTube Web page explaining the program, where the company's position on the accuracy of the program is also stated. "No matter how accurate the tools get, it is important to remember that no technology can tell legal from infringing material without the cooperation of the content owners themselves," according to information on the Web site written jointly by YouTube's King and Glenn Brown, partner development manager. Meanwhile, consumer advocacy groups may begin to worry that the new YouTube Video Identification System will block legitimately posted video content, including those uploads that comply with copyright's "fair use" exception. Where Does the Burden of Identifying Content Lie? Should YouTube's Video Identification system prove successful it may help in sealing the rift that has developed over the years between online video sharing services and media companies. Several large media companies have been working on their own program aimed at stamping out illegal content sharing, and later this week an announcement is expected from a group including Viacom, Disney and Microsoft detailing a cooperative plan to overcome the problem. YouTube notes that their new system is aimed at a broad spectrum of people. "As we scale and refine our system, YouTube Video Identification will be available to all kinds of copyright holders all over the world, whether they want their content to appear on YouTube or not," according to the company Web site. YouTube's "Brand-new, cutting edge" Video Copyright System The technology behind using so-called "digital fingerprinting" is not new to Google, who has filed several patents over the past years that deal with the identification of both video and audio content. The first patent was filed in November of 2006 and published on June 7, 2007, and is entitled “Social and Interactive Applications for Mass Media.” The patent covers a scenario where Google could use devices containing a microphone or video camera placed in consumer’s homes to inform them which television show, radio show or movie the consumer is viewing or listening to, and presenting related material including targeted advertisements to the consumer through their computers, cell phones, or wireless devices. This would involve Google maintaining a database of television programs, radio programs, and films, and to check the information received from the device in the consumer’s residence in order to determine what is being watched or listened to, and then providing “customized and interactive information” to the consumer. The patent identified this information as including real-time popularity ratings, video and audio book marking, personalized information layers, and social peer communities of consumers watching or listening to the same programs. This previous patent technology would allow Google to determine such things as when a consumer is viewing or listening to a previously recorded advertisement, and presenting a new advertisement in its place using the consumer’s computer, cell phone or wireless device. The patents deal with several processes used to verify what is being viewed or listened to, and also with identifying what comes before and after any known television or radio program. Google's "Audioswap" program is also mentioned on the YouTube Web page announcing the new video identification program, and offers further evidence that the company has been involved in digital fingerprinting for some time. Google is not the only large Internet company to develop such identification programs. Both Microsoft and MySpace also use digital fingerprinting technology to find and stop the uploading of copyrighted material. In June of this year, Google first commented on their video identification tools: We’ve been developing improved content identification for months, and we’re confident that in the not-too-distant future, we’ll unveil an innovative solution that will work for users and content creators alike. This is one of the most technologically complicated tasks that we have ever undertaken. But YouTube has always been committed to developing sustainable and scalable tools that work for all content owners. That is what YouTube co-founder Steve Chen wrote on the official Google blog at the time. Yesterday's announcement of the new YouTube Video Identification system will likely play a large role in shaping the future of online video sharing. ¹ According to Miguel Helft's October 16, 2007 New York Times article citing "people with knowledge of the tests." Related Links:
| 
10:57 am on Oct. 16, 2007 (utc 0)