Google Takes Step on Video Copyrights

By MIGUEL HELFT

Published: October 16, 2007

SAN BRUNO, Calif., Oct. 15 — Google is seeking to put an end to the copyright wars over online video.

On Monday, the company unveiled a long-anticipated system that, if effective, would allow media companies to prevent their clips from being uploaded to YouTube without permission.

Whether the system will work well enough to satisfy media companies who have been irked by the proliferation of unauthorized copyrighted clips on YouTube is not yet clear. But if successful, the system, which Google is offering to all media companies, could usher in a détente between them and Google.

“We are delighted that Google appears to be stepping up to its responsibility and ending the practice of profiting from infringement,” said Michael Fricklas, general counsel of Viacom, which filed a $1 billion copyright infringement suit against YouTube and Google in March. “We’ll be watching to ensure that the system is reasonably effective and sufficiently robust to address the issue.”

Google said it had been testing the system with nine media companies, including Time Warner, CBS and Disney. Others involved include NBC Universal and Viacom, according to people with knowledge of the tests.

Google called the tests “promising” but would not say how effective the system was. Just last week, its chief executive, Eric E. Schmidt, said that developing a system that could identify video clips with 100 percent accuracy was virtually impossible.

“The question is, Can we get to 80 or 90 percent?” Mr. Schmidt said in an interview with a group of reporters.

It is not known if the system has reached that level of accuracy. At least one of Google’s testing partners said that it was a work in progress.

“They still have a ways to go with the system before we could call it totally sufficient,” said Edward Adler, executive vice president for communications at Time Warner.

Google said that its video identification service, which was developed by its own engineers, required media companies to submit their digital video files to Google, which would then create what technologists call a digital “fingerprint” for each file. That fingerprint would then be uploaded to a large database. Once a user uploaded a new clip, the same technology would determine whether that clip’s fingerprint matched a fingerprint in the database.

Content owners could instruct Google to block clips whose fingerprints matched their copyrighted clips. Alternatively, they could ask Google to promote the clip and even place advertising around it, to share revenue from the ads.

“We really need the content community to work with us,” said David King, a YouTube product manager. “What really drives this whole thing is having access to the reference material.”

Clips that content owners want blocked may be posted on YouTube for a few minutes before they are taken down, but over time, the company hopes to speed up the identification mechanism so that unauthorized uploads can be prevented altogether.

YouTube representatives said the system was able to identify clips that were identical to those in its database, as well as those that had been slightly modified by users to escape detection. To demonstrate, they showed a positive identification of a clip that had been captured by a camera filming a television set as the clip was being shown.

Google has been using fingerprints to recognize audio files for some time. Others, like MySpace and Microsoft, also use audio fingerprints to identify, and sometimes block, uploads of certain videos.

Google introduced the video identification system as a group of media and technology companies, including Microsoft, Viacom, Disney and others, are set to unveil a framework for how the two sides should cooperate to stamp out copyright infringement online, said a senior executive at a media company.

The executive said the group would make public that framework later this week and that, along with Google’s announcement Monday, it represented a “potential recipe for working relationships between content companies and digital distributors of all kinds.”

But not all media companies agree that Google’s identification system is enough.

“I think this is a completely inadequate solution,” said Louis Solomon, a partner in Proskauer Rose, which represents the Football Association Premier League of England, a lead plaintiff in a class-action copyright suit against Google. “It is too late in coming; it offers too little protection; it gives YouTube and Google content that they don’t need and shouldn’t have.”

Some consumer groups, meanwhile, worry that Google’s new system could prevent uploads of video clips that were authorized under “fair use” provisions of copyright law.

For its part, Viacom said it was too early to say what impact Google’s new system would have on its suit.

“We obviously have suffered significant past damages, but beyond that it is premature to tell what the effect would be on the litigation,” Mr. Fricklas said.

Google has long insisted that YouTube has always operated in compliance with copyright law, in part because it takes down unauthorized copyrighted clips when asked to by content owners. The new system would help content owners by automating that process, the company said.