Nabbing Video Pirates: Who Needs Google?

YouTube is working on software to spot copyrighted postings, but content owners may not wait for—or want—the system it has in mind

by Peter Burrows

But will it work? It's a question likely to be asked by a lot of would-be users of new tools created by Google (GOOG) that are designed to ferret out video posted illegally to the search giant's YouTube site.

The answer is critical to YouTube, which unveiled the content-recognition technology Oct. 15 at its San Bruno (Calif.) headquarters, as well as to the owners of copyrighted movies, TV shows, and other content that is routinely uploaded and viewed on the site and others like it. While news networks like General Electric's (GE) NBC, movie studios including News Corp.'s (NWS) 20th Century Fox, and others are racing to put their content online to reach the legions of people who have flocked there, they've been looking for ways to make sure that content isn't pirated. Theoretically, these new tools will enable video sites to track such content, and give studios and other video creators a choice on what to do when pirated clips appear. Content owners could insist that the clips be removed from a site, allow them to be viewed as a promotional vehicle, or even let the clips be shown with ads—creating revenue-sharing opportunities between the owner and the site.

The challenge to devise a better content-recognition system has touched off a race among companies—from startups like Vobile to giants like Philips (PHG)—to bring technologies to market. A dozen took part in a trial held by the Motion Picture Association of America that ended earlier this fall. YouTube, which was notably absent from the trial, has been developing its own.

The pressure is on Google to come up with a better way to keep copyrighted programming off YouTube, but critics have questioned the company's commitment to the effort. Industry sources have expressed frustration with YouTube's unwillingness to join the MPAA trial and say the company has a financial interest in dragging its feet in complying with content takedown orders. In March, Viacom (VIA) sued Google, accusing YouTube of massive copyright infringement (BusinessWeek.com, 3/14/07).

"Firmly Aligned Behind the Goal"

YouTube representatives say the effort to build content-recognition software undermines those criticisms. "We are firmly aligned behind the goal" of copyright protection, product manager David King said during the presentation. He pointed out various moves taken by the company in the past, such as using what's called a "hash" technology that can find perfect, unaltered copies of a clip, and the use of technology from Audible Magic to spot pirated audio soundtracks.

In the Oct. 15 press conference, YouTube provided the basics of how the technology would work. Like some of the other systems, it depends on content owners handing over their content to the company, which would then assemble a database of "digital IDs." Then, whenever someone uploads or tries to view a video, a "video identification engine" quickly scans it to look for matches. Once a pirated clip is spotted, the content owner has three choices: leave it up, block it (meaning it won't show up in search results as well), or monetize it. Currently, this means sharing in Google revenue from ads that can be placed around the clip.

There are a number of potential problems. For starters, will it work? According to King, the product has been used for only a week on actual content, from a single video partner. The system found just 18 pirated clips in that time—a tiny number given the vast trade in copyrighted material on the Web.