<i>Perfect 10 v. Google</i>: Adult Web Site Challenges Search Engine On Image Hits

Could the right to search copyrighted images on the Internet be in jeopardy? That would be outcome if Perfect 10, Inc. (P10), a purveyor of adult entertainment, has its way.

P10's main grievance is that obscure Web sites, which are often judgment-proof and/or hosted in foreign countries, have posted copies of copyrighted P10 photos that “easily can be located en masse” through Google, Inc., the Goliath search engine. Its solution: Sue Google.

Late last year, P10 filed a complaint in federal court in California against Google alleging claims of direct, contributory and vicarious copyright infringement, trademark infringement, and violation of California's infamous Bus. & Prof. Code ' 17200, among others. [Editor's Note: Our December issue's Net News included an item on the filing of the lawsuit.]

An outcome in P10's favor would have a seismic impact upon the operation of Internet search engines and, therefore, the utility of the Internet itself.

A Recurrent Litigant

Perfect 10, owned by Norman Zadeh, sells an adult magazine and operates a fee-based Web site; few have heard of either. P10 and Zadeh achieved some success as a crusader against sites offering scans of copyrighted images from Playboy and other magazines. Zadeh's idea was not to attack the Webmasters themselves ' a fruitless exercise ' but rather to go after the infrastructure supporting them.

In 2002, P10 filed a copyright, trademark and unfair competition case against the credit card-based age verification services (AVS) used by adult Web sites to keep minors out.

AVSs, such as Cybernet's Adult Check, charge a fee for each password, and pass back a portion to the “referring” site. P10's theory was that the AVS had reason to know that the referring sites were populated with thousands of infringing scans. By making money from infringing content, Zadeh alleged the AVS firms should be held liable for contributory or vicarious copyright infringement. P10 won a preliminary injunction in the case against the AVS, and the case settled shortly thereafter.

Next, P10 took on the processors of credit card charges used by adult Web sites on basically the same theory. P10 had a mixed record resisting motions for summary judgment in this litigation, and that case has also appears to have settled.

P10 next turned its sights on Visa and Mastercard, alleging that each was a “virtual fence” for the infringing Web sites. The Visa/Mastercard suit did not get very far. On a pre-answer motion to dismiss, the court had little trouble concluding that processing credit card charges was too remote from actual infringing content for the defendants to be liable. P10 was also required to pay hundreds of thousands of dollars in attorney's fees for the prevailing defendants, but plans to appeal.

The P10 Google Complaint

Undaunted, P10 has now gone in search of another deep pocket.

In Google, P10 chose perhaps the ultimate deep pocket on the Internet. P10's complaint alleges that its site “receives approximately 100,000 unique visitors per month,” some of which ' if that allegation is true ' find its site via Google.

In P10's complaint, it lists Google along with “Doe” Web site operators that, P10 alleges, have “acted with defendant Google and is responsible for the harm and damages [caused to Perfect 10].”

These anonymous Doe defendants are alleged to operate sites ' dubbed “stolen content” Web sites (SCWs) ' with unauthorized or unlicensed content, eg, infringing copies of P10 images. P10 complains that “because many of the [SCWs] are judgment proof, it is economically and practically impossible to sue them for infringement.”

P10 also states that “[i]t would be virtually impossible for consumers to locate most [SCWs] if they were not directed to them by [Google].”

Additionally, because the SCWs sometimes display images with “visible Perfect 10 marks or a visible copyright notice,” P10 alleges porn aficionados are “mislead … into believing [the SCWs] are associated, affiliated with or authorized by Perfect 10.”

P10 also claims that Google, through its Adwords and Adsense programs, has set up means by which the sites with infringing content “can, for a fee, have their advertisements appear in response to searches performed on various key words, including those constituting the Perfect 10 marks and protected by the Perfect 10 Rights of Publicity.”

The complaint alleges that members of Google's Adwords and Adsense programs receive preferential treatment in that “[w]hen a consumer performs a search performed on [various] keywords, the website will appear at or near the top of the generated list of search results provided by [Google].”

P10 alleges that Google provides access to consumers of P10's copyrighted works from SCWs via Google's search functions: a) the “Web search,” the standard search by which terms are inserted and a list of links in order of relevance, including a description of the content with the terms requested; and b) “Search Images” where a term search can allegedly “yield high-quality copies of images, including thousands of copies of the Perfect 10 copyrighted works.” Specifically, the complaint alleges that this infringement is accomplished (directly, contributory and vicariously) by Google in at least five ways, all of which generate visitors and revenue to Google:

1. Copying P10's copyrighted works to Google's server and providing the images to consumers in two different sizes through its “search images function”;
2. Displaying what appears to be a “full sized” P10 image in a separate window so that a consumer believes it is actually located on Google's server;
3. Creating multiple links to SCWs through Google's Web search function so that when a search is performed on a P10 model's name, “the links which appear at the top of the search results frequently lead to free, full sized images of that model”;
4. Linking consumers to sites featuring hacked P10 passwords for free access to the Web site; and
5. “[P]urport[ing] to market and sell the right to use P10 marks and P10 rights of publicity to advertisers and entities, including [infringing sites], which defendants know have no connection to Perfect 10.”

P10 also alleges to have provided Google actual notice of these infringements in the form of at least 27 Digital Millennium Copyright Act (DMCA)-compliant notices. These DMCA notices allegedly informed Google of infringements on at least 2,700 URLs in its system, including approximately 1000 URLs for infringing images on its servers.

On this basis, P10 concludes that, as of November 2004, there were at least 8 million links in Google's system, “leading to thousands of distinct copies of unauthorized P10 copyrighted works” on sites to which Google links.

Does P10 Stand a Chance?

Given the current state of the law on image search engines, P10's chances of success appear minimal. Search engines such as Google and others like it are typically composed of three parts: a) the “spider” that crawls the Web in order to create a list of “hits”; b) the list of “hits” is commonly known as the search engine's “index”; and c) search and retrieval software.

The effectiveness of a term-based search depends on the comprehensiveness of the search engine's database, the sophistication of its search and retrieval software, and the users' skill in crafting an appropriate query.

Using Google's “search images” function, Google retrieves pictures meeting the parameters of the search, displayed as an array of hyperlinked “thumbnail” images. Clicking on an image results in the top quarter of the page remaining on Google's site and the bottom frame displaying the content at the linking site retrieved from the search.

This is known as hyperlink framing, which allows a Web page designer to link to another's site while still maintaining the “frame” of their home site. The material within the linked frame will be presented exactly as if the user had accessed the URL address of that Web page directly.

Hyperlinks are essential to the existence of the Internet. Commentators have called them “the threads with which the Web is spun.”

The main advantage of linking is convenience. Links make the Internet a powerful resource that provides its users with an interconnected world of knowledge, all made possible by the ability to link information available on one site to other sites.

The linking of information into a “single body of knowledge,” which is in turn utilized by search engines such as Google, is central to the promotion of the free exchange of ideas and information.

In fact, commentators and courts alike have often extolled search engines and the benefits derived from their ease of use.

Search engines retrieve images regardless of whether the images were lawfully displayed by the responding site. Search engine computers are essentially “dumb” to copyright and trademark issues, as one would expect. For this reason, and given the salutary purposes served by search engines, courts have been loathe to find violations of the Copyright Act for fundamentally necessary functions of the Web, such as hyperlinking.

Arriba Soft

The leading case on image search engines is Kelly v. Arriba Soft Corporation, 336 F.3d 811, 820 (9th Cir. 2003).

In Arriba Soft, the defendant operated a search engine that crawled the Web, copying images and creating thumbnail-sized reproductions of those images. The plaintiff, a professional photographer who operated a site with his copyrighted images of the American West, sued the search engine alleging copyright claims.

After finding that the search engine operated its site for obvious commercial purposes because of the advertising surrounding the images, the U.S. Court of Appeals for the Ninth Circuit held that the search engine's creation and display of thumbnail-sized images was a fair use under the Copyright Act.

In so doing, the court noted that the search engine's use of the copyrighted images as thumbnails was transformative of the nature, purpose and character of the purpose the images served on the plaintiff's Web site.

Turning art into news or information into art is transformative of the original work's purpose and indicative of fair use.

Google's image search engine stands in good stead under the Arriba Soft standard. Google displays a page full of thumbnails with no “sponsored links,” as contrasted with the results of text searches. The thumbnails, which under Arriba Soft should be considered “fair use” in a search engine context, are further insulated from liability because they are the hyperlink. The Web address beneath each thumbnail is not hypertext. So Google should get the benefit of holdings such as that in Ticketmaster Corp. v. Tickets.com, Inc., 2000 WL 525390 (C.D. Cal. 2000), which extolled the benefits of hyperlinking:

[H]yperlinking does not itself involve a violation of the Copyright act … since no copying is involved. The customer is automatically transferred to the particular genuine Web page of the original author. There is no deception in what is happening. This is analogous to using a library's card index to get reference to particular items, albeit faster and more efficiently.

Where a search engine is alleged to infringe a copyright holder's exclusive right to reproduce, distribute, and publicly display copies of the work, a prima facie case of copyright infringement will lie but courts have authority to avoid “rigid application of the copyright statute … [where it] would stifle the very creativity which that law is designed to foster.”

Could the right to search copyrighted images on the Internet be in jeopardy? That would be outcome if Perfect 10, Inc. (P10), a purveyor of adult entertainment, has its way.

P10's main grievance is that obscure Web sites, which are often judgment-proof and/or hosted in foreign countries, have posted copies of copyrighted P10 photos that “easily can be located en masse” through Google, Inc., the Goliath search engine. Its solution: Sue Google.

Late last year, P10 filed a complaint in federal court in California against Google alleging claims of direct, contributory and vicarious copyright infringement, trademark infringement, and violation of California's infamous Bus. & Prof. Code ' 17200, among others. [Editor's Note: Our December issue's Net News included an item on the filing of the lawsuit.]

An outcome in P10's favor would have a seismic impact upon the operation of Internet search engines and, therefore, the utility of the Internet itself.

A Recurrent Litigant

Perfect 10, owned by Norman Zadeh, sells an adult magazine and operates a fee-based Web site; few have heard of either. P10 and Zadeh achieved some success as a crusader against sites offering scans of copyrighted images from Playboy and other magazines. Zadeh's idea was not to attack the Webmasters themselves ' a fruitless exercise ' but rather to go after the infrastructure supporting them.

In 2002, P10 filed a copyright, trademark and unfair competition case against the credit card-based age verification services (AVS) used by adult Web sites to keep minors out.

AVSs, such as Cybernet's Adult Check, charge a fee for each password, and pass back a portion to the “referring” site. P10's theory was that the AVS had reason to know that the referring sites were populated with thousands of infringing scans. By making money from infringing content, Zadeh alleged the AVS firms should be held liable for contributory or vicarious copyright infringement. P10 won a preliminary injunction in the case against the AVS, and the case settled shortly thereafter.

Next, P10 took on the processors of credit card charges used by adult Web sites on basically the same theory. P10 had a mixed record resisting motions for summary judgment in this litigation, and that case has also appears to have settled.

P10 next turned its sights on Visa and Mastercard, alleging that each was a “virtual fence” for the infringing Web sites. The Visa/Mastercard suit did not get very far. On a pre-answer motion to dismiss, the court had little trouble concluding that processing credit card charges was too remote from actual infringing content for the defendants to be liable. P10 was also required to pay hundreds of thousands of dollars in attorney's fees for the prevailing defendants, but plans to appeal.

The P10 Google Complaint

Undaunted, P10 has now gone in search of another deep pocket.

In Google, P10 chose perhaps the ultimate deep pocket on the Internet. P10's complaint alleges that its site “receives approximately 100,000 unique visitors per month,” some of which ' if that allegation is true ' find its site via Google.

In P10's complaint, it lists Google along with “Doe” Web site operators that, P10 alleges, have “acted with defendant Google and is responsible for the harm and damages [caused to Perfect 10].”

These anonymous Doe defendants are alleged to operate sites ' dubbed “stolen content” Web sites (SCWs) ' with unauthorized or unlicensed content, eg, infringing copies of P10 images. P10 complains that “because many of the [SCWs] are judgment proof, it is economically and practically impossible to sue them for infringement.”

P10 also states that “[i]t would be virtually impossible for consumers to locate most [SCWs] if they were not directed to them by [Google].”

Additionally, because the SCWs sometimes display images with “visible Perfect 10 marks or a visible copyright notice,” P10 alleges porn aficionados are “mislead … into believing [the SCWs] are associated, affiliated with or authorized by Perfect 10.”

P10 also claims that Google, through its Adwords and Adsense programs, has set up means by which the sites with infringing content “can, for a fee, have their advertisements appear in response to searches performed on various key words, including those constituting the Perfect 10 marks and protected by the Perfect 10 Rights of Publicity.”

The complaint alleges that members of Google's Adwords and Adsense programs receive preferential treatment in that “[w]hen a consumer performs a search performed on [various] keywords, the website will appear at or near the top of the generated list of search results provided by [Google].”

P10 alleges that Google provides access to consumers of P10's copyrighted works from SCWs via Google's search functions: a) the “Web search,” the standard search by which terms are inserted and a list of links in order of relevance, including a description of the content with the terms requested; and b) “Search Images” where a term search can allegedly “yield high-quality copies of images, including thousands of copies of the Perfect 10 copyrighted works.” Specifically, the complaint alleges that this infringement is accomplished (directly, contributory and vicariously) by Google in at least five ways, all of which generate visitors and revenue to Google:

1. Copying P10's copyrighted works to Google's server and providing the images to consumers in two different sizes through its “search images function”;
2. Displaying what appears to be a “full sized” P10 image in a separate window so that a consumer believes it is actually located on Google's server;
3. Creating multiple links to SCWs through Google's Web search function so that when a search is performed on a P10 model's name, “the links which appear at the top of the search results frequently lead to free, full sized images of that model”;
4. Linking consumers to sites featuring hacked P10 passwords for free access to the Web site; and
5. “[P]urport[ing] to market and sell the right to use P10 marks and P10 rights of publicity to advertisers and entities, including [infringing sites], which defendants know have no connection to Perfect 10.”

P10 also alleges to have provided Google actual notice of these infringements in the form of at least 27 Digital Millennium Copyright Act (DMCA)-compliant notices. These DMCA notices allegedly informed Google of infringements on at least 2,700 URLs in its system, including approximately 1000 URLs for infringing images on its servers.

On this basis, P10 concludes that, as of November 2004, there were at least 8 million links in Google's system, “leading to thousands of distinct copies of unauthorized P10 copyrighted works” on sites to which Google links.

Does P10 Stand a Chance?

Given the current state of the law on image search engines, P10's chances of success appear minimal. Search engines such as Google and others like it are typically composed of three parts: a) the “spider” that crawls the Web in order to create a list of “hits”; b) the list of “hits” is commonly known as the search engine's “index”; and c) search and retrieval software.

The effectiveness of a term-based search depends on the comprehensiveness of the search engine's database, the sophistication of its search and retrieval software, and the users' skill in crafting an appropriate query.

Using Google's “search images” function, Google retrieves pictures meeting the parameters of the search, displayed as an array of hyperlinked “thumbnail” images. Clicking on an image results in the top quarter of the page remaining on Google's site and the bottom frame displaying the content at the linking site retrieved from the search.

This is known as hyperlink framing, which allows a Web page designer to link to another's site while still maintaining the “frame” of their home site. The material within the linked frame will be presented exactly as if the user had accessed the URL address of that Web page directly.

Hyperlinks are essential to the existence of the Internet. Commentators have called them “the threads with which the Web is spun.”

The main advantage of linking is convenience. Links make the Internet a powerful resource that provides its users with an interconnected world of knowledge, all made possible by the ability to link information available on one site to other sites.

The linking of information into a “single body of knowledge,” which is in turn utilized by search engines such as Google, is central to the promotion of the free exchange of ideas and information.

In fact, commentators and courts alike have often extolled search engines and the benefits derived from their ease of use.

Search engines retrieve images regardless of whether the images were lawfully displayed by the responding site. Search engine computers are essentially “dumb” to copyright and trademark issues, as one would expect. For this reason, and given the salutary purposes served by search engines, courts have been loathe to find violations of the Copyright Act for fundamentally necessary functions of the Web, such as hyperlinking.

Arriba Soft

The leading case on image search engines is Kelly v. Arriba Soft Corporation , 336 F.3d 811, 820 (9th Cir. 2003).

In Arriba Soft, the defendant operated a search engine that crawled the Web, copying images and creating thumbnail-sized reproductions of those images. The plaintiff, a professional photographer who operated a site with his copyrighted images of the American West, sued the search engine alleging copyright claims.

After finding that the search engine operated its site for obvious commercial purposes because of the advertising surrounding the images, the U.S. Court of Appeals for the Ninth Circuit held that the search engine's creation and display of thumbnail-sized images was a fair use under the Copyright Act.

In so doing, the court noted that the search engine's use of the copyrighted images as thumbnails was transformative of the nature, purpose and character of the purpose the images served on the plaintiff's Web site.

Turning art into news or information into art is transformative of the original work's purpose and indicative of fair use.

Google's image search engine stands in good stead under the Arriba Soft standard. Google displays a page full of thumbnails with no “sponsored links,” as contrasted with the results of text searches. The thumbnails, which under Arriba Soft should be considered “fair use” in a search engine context, are further insulated from liability because they are the hyperlink. The Web address beneath each thumbnail is not hypertext. So Google should get the benefit of holdings such as that in Ticketmaster Corp. v. Tickets.com, Inc., 2000 WL 525390 (C.D. Cal. 2000), which extolled the benefits of hyperlinking:

[H]yperlinking does not itself involve a violation of the Copyright act … since no copying is involved. The customer is automatically transferred to the particular genuine Web page of the original author. There is no deception in what is happening. This is analogous to using a library's card index to get reference to particular items, albeit faster and more efficiently.

Where a search engine is alleged to infringe a copyright holder's exclusive right to reproduce, distribute, and publicly display copies of the work, a prima facie case of copyright infringement will lie but courts have authority to avoid “rigid application of the copyright statute … [where it] would stifle the very creativity which that law is designed to foster.”