Legal Analysis by Corynne McSherry
Copyright owners, take note: If you’re going to use the streamlined Digital Millennium Copyright Act (“DMCA”) process to require a service provider to remove allegedly infringing content, you’d better make sure you actually comply with the DMCA notice requirements. Otherwise a court may find, as occurred this week in Perfect 10 v. Google, that your “notice” didn’t actually put anyone on “notice.”
A quick recap: In 2004, porn company and frequent litigator Perfect 10 sued Google for direct and secondary copyright infringement. Perfect 10 claimed that Google violated its copyrights by linking to websites that hosted infringing material, caching websites that hosted infringing photos of nude models, and hosting infringing images uploaded by Blogger users. In 2007, the Ninth Circuit Court of Appeals reversed a preliminary injunction in favor of Perfect 10 on its direct infringement claims, and sent the case back to the district court for a determination of some of the secondary infringement claims. Google moved for summary judgment, asserting that it was protected from secondary liability by the DMCA safe harbors.
This week, Judge Howard Matz of the U.S. District Court in Los Angeles mostly agreed with Google, whittling Perfect 10’s remaining case down to a small subset of allegedly infringing images. Why? Mainly because Perfect 10 didn’t trouble itself to provide Google with the information Google needed to figure out what to take down in a form that Google could readily use.
The DMCA requires a proper takedown notice to identify the work claimed to be infringed, identify the reference (or link) to material claimed to be infringing, and provide enough information to permit the service provider to locate that reference or link. Even though providing this information should be pretty easy, Perfect 10 fell far short.
For example, many of its “notices” consisted of a cover letter, a spreadsheet with URLs (many of which linked only to a top-level URL for a website, as opposed to a specific infringing URL) and a hard drive or DVD containing Perfect 10’s electronic files of its photos. Not good enough, said the court — the information required by the DMCA must be contained in a single written communication; forcing a service provider to cobble together adequate notice from a variety of sources is just too burdensome.
P10 evidently expected Google to comb through hundreds of nested electronic folders containing over 70,000 distinct files, including raw image files such as JPEG files and screen shots of Google search results, in order to find which link was allegedly infringing.  In many cases, the file containing the allegedly infringing image does not even include a URL, or the URL was truncated.  The spreadsheets also do not identify the copyrighted work that was allegedly infringed. . . . P10 then expected Google to search through a separate electronic folder—attached only to the June 28, 2007 DMCA notice—containing all of the more than 15,000 images that appeared on P10’s website as of June 2007, in order to identify the copyrighted work that was infringed.
The court did find that a small subset of notices complied with the DMCA; for those few notices, Google must now show that it responded to those notices “expeditiously” under the circumstances.
The ruling is not unprecedented: numerous courts, including the Ninth Circuit have found that ISPs don’t have to respond to deficient DMCA notices. But the issue of how much information about infringement providers need to have (and fail to act on) before they lose the protection of the safe harbors is being hotly contested in two other proceedings. Content owners are insisting that if ISPs have general knowledge of infringement on their services, they must take over the burden of stamping it out. We might think of Judge Matz’s decision as one more vote in favor of keeping the burden of identifying copyright infringement where it has traditionally belonged — on the content owners themselves.