Mass Digitization

While Congress may have sidelined itself, not so the lower and appellate courts. Mass digitization projects contribute to the pressure on business models to adapt to the "attention" economy by offering ease of access and freedom to use products and services as the consumer desires. But the massive push to put everything online has indirectly changed the law because such massive amounts of online information require effective means for their organization and access. The most successful method so far is the search engine. But search engines carry out their functions by duplicating in their entirety the works to be placed in the index, and displaying and distributing links to the works to be served as search results. Plaintiffs have challenged these and other features of search and, given the value search engines add to the masses of online information, we should not be surprised that the courts have refused to interpret the law as putting search engine companies in the untenable position of having to ask permission before indexing and serving up information about the existence and description of others' works.

Conceivably, Congress could create a search engine exception to infringement liability. Indeed, Gigi Sohn of Public Knowledge called for just such an amendment at the 2007 New Media and the Marketplace of Ideas Conference, but that is not likely to happen (as Gigi no doubt knows). So the action is taking place in court. The search engine cases elaborate the scope of the only existing exception capable of handling the situation, and are thereby giving new definition to the scope of fair use and contributing to the ongoing policy debate about the scope of secondary liability for others' infringements. This series of cases began with Kelly v. Arriba Soft, continues with Field v. Google and Perfect 10 v. Amazon & Google. All of these involve search engines making and displaying copies of plaintiffs' works as part of normal search engine operation. The fair use analysis in each case is similar in that the critical aspects include first, characterizing search engine use of others' works as highly transformative, and then weighing and balancing the social benefit of the search function against the harm, if any, to the copyright owner. The details vary from case to case, of course, but the pattern has been established. If you place materials online, they will be crawled, your work will be copied, contents indexed, thumbnails of images created, and the works included in results pages, with links to the originals, in accordance with the relevance and other criteria built into the search engine's algorithms. If a copyright owner does not want this to happen to her online material, she indicates her desire that Web crawlers ignore her work (she opts out) by using the robots.txt tag. End of discussion.

It sounds easy now that it's been fairly well-established, but there were several major hurdles along the way. The courts had to characterize thumbnails (reduced, lower-resolution versions of full-sized images, as defined in Perfect 10 v. Amazon, p. 5761) as fair use, establish that caching was fair use (notably in Field v. Google), and establish the relative social value of search engines compared to the possible harm to a plaintiff's market interests (explored in all three cases). As mentioned above, since in all cases the decisive facts about the infringing use were the highly transformative nature and great social value of the search engine's copying and displaying, we can surmise that such massive infringement in the service of a goal or aim that merely competed with the plaintiff's purpose would not have prevailed. The search engines have a wholly different purpose from the purpose underlying content providers' creation and distribution, however -- to help us make sense of the billions of works on the Internet. And the courts recognize that important purpose and are not inclined so far to underestimate its value in comparison to the value of the economic incentive to copyright owners. That's the essence of fair use: balancing competing interests. Even if there were some harm or potential for harm to plaintiff's market interests, that harm or potential for harm must be sufficiently serious to outweigh the social benefit provided by search engines. Thus, it may be that a copyright owner could sell little tiny pictures (thumbnails) of naked women to cell phone owners, but if the court must choose between privileging those possible sales and requiring search engines to do the impossible, that is, ask permission from every Web page owner before crawling a page -- that has not proved to be a tough choice.

So, search functionality, this very important and necessary service that organizes and provides access to the massive amounts of digital materials available on the Web today, has indeed affected copyright law. The search engine cases refined the application of the function of fair use enabling transformative uses. The Perfect 10 v. Amazon & Google case is poised to further refine the law of secondary contributory liability as well. The 9th circuit court sent the matter back to the district court for further fact finding on the issue of whether Google's linking to infringing images on third parties' servers contributes to infringing activity. The district court originally held that it did not, but the 9th circuit disagreed. Even if Google does contribute to infringement by linking to infringing copies, it claims protection of the Internet service provider liability limitations in Section 512, those provisions that protect universities as well as major Internet service providers who provide information location tools (links to others' content). The dispute will center on what steps the law requires that Google take in response to a notice from Perfect 10 that its images are being infringed. Perfect 10 will argue that Google should do more than it has done so far. Google will argue that what it has done is all the law requires, that is, respond to notices of alleged infringement.