See Digital Millennium Copyright Act, Title II


A search engine is a computer program that locates files stored on one or more computers. The user specifies values, such as words, numbers, or dates, for the files sought; the search engine then examines the files in the location it is designed to search and returns a list of files containing those values. There are a wide variety of search engines for different tasks. The Windows Search function is a search engine that searches a single personal computer. Library online catalogs search a network of library computers, and may connect to search engines at other libraries or search a shared inter-library database. Search engines make P2P networks possible. Search engines for large proprietary professional databases such as Westlaw and Lexis offer sophisticated search functions that require training to use. But the search engines that attract the greatest amount of attention, and are most likely to raise legal issues, are those that search the World Wide Web. Some of these search only a portion of the Web, while others attempt to search the entire Internet, and actually do search a large portion of it. The most popular of these general-purpose search engines is Google, followed by Microsoft’s Bing, Yahoo! Search, and Ask; Google, however, has more unique monthly visitors than all of its competitors combined.

Internet search engines may become involved in legal disputes in several ways. They may be vehicles for copyright and trademark infringement. Because they may be used from anywhere in the world, they may violate local censorship laws. And they are subject to abuse by search engine spammers.

Search engines and copyright infringement

Internet search engines are generally considered to be service providers within the meaning of Title II of the Digital Millennium Copyright Act (DMCA), although Title II as a whole seems not to have been designed with search engines in mind: references to “subscribers” and “account holders” have no application to search engines, for example. As service providers, search engines are eligible for the protection of the four statutory safe havens from liability for copyright infringement: the transitory communications safe harbor, the system caching safe harbor, the information storage safe harbor, and the information location tools safe harbor. The fourth of these, the information location tools safe harbor, is the one designed with search engines in mind; however, the nature of most search engines means that some search engine activities may also fall within the other three safe harbors. Most Internet search engines, including the most popular ones, work by crawling the Web and storing information about each crawled Web site in the search engine’s database. Searches then take place in the database rather than in the Web as a whole, which makes the searches much faster. Some, such as Google, also make cached versions of Web sites available to the public.

The transitory communications safe harbor shields a service provider from liability where the provider merely acts as a conduit, transmitting digital information from one point on a network to another at someone else’s request. Five elements must be present for a service provider to avail itself of the transitory communications safe harbor: the transmission must be “initiated by or at the direction of a person other than the” service provider; it must be “carried out through an automatic technical process without selection of the material by the” service provider; the service provider must “not select the recipients of the material except as an automatic response to the request of another person”; the service provider must not maintain a copy of the material “in a manner ordinarily accessible to anyone other than anticipated recipients” or “for a longer period than is reasonably necessary”; and the material must be transmitted “without modification of its content” (17 U.S.C. § 512(a)).

Many search engine activities are likely to meet these requirements: the search engine can be viewed as a conduit for conveying information about Web sites between Web site operators and users. The users initiate the search, which is carried out by an automatic technical process, although it might be difficult to establish that the process does not involve selection of the material by the service provider. The service provider does not select the recipients of the material; the recipients select themselves. The results of a particular search are accessible only to the person running that search, and are maintained only until the recipient closes the search window or runs another search. However, there are problems with this reasoning; it may well be that the automated process of selecting search results is “selection of material” within the meaning of the statute. The transitory communications safe harbor is not generally applied to search engine activities.

The system caching safe harbor shields a service provider from liability where the provider makes a copy of material for system caching purposes. This system caching is a necessary part of the operation of all search engines that rely on searching their own databases of crawled Web sites, which is to say nearly all Internet search engines. In order for a service provider to be eligible for the system caching safe harbor, three requirements must be met: the material must be “made available online by a person other than the service provider”; it must be transmitted from that person to another person at the direction of the second person; and the caching “must be carried out through an automatic technical process for the purpose of making the material available to users of the system or network who . . . request access” (17 U.S.C. § 512(b)). These conditions appear to be met in the case of Internet search engines, although again the poorness of the fit between the statute and the reality of search engine operation is evident. Information about Web pages is made available online by the Web site operators, not by the search engines. It is “transmitted,” or at least communicated in some way, to search engine users at their request. And the caching is done by an automatic process for the purpose of making the information available to search engine users.

Even if these requirements are met and the search engine service meets the threshold requirement of eligibility for the system caching safe harbor, however, additional conditions must be met before the search engine is shielded from liability. The search engine must not modify the retained material, must periodically refresh it, and must not interfere with technology that returns hit information to the person who posted the material. Search engines are likely to meet these conditions, but an additional condition has caused some controversy: once the service provider has been notified that copyrighted material posted without authorization has been removed or blocked, or ordered to be removed or blocked, at the original site, the service provider must promptly remove or block access to the cached copy. If the person who posted the material has imposed limitations on access, such as password protection, the service provider must also impose these limitations (17 U.S.C. § 512(b)(2)). This provision may be necessary to protect the interests of the copyright owner, but it interferes with the use of large cached databases such as Google’s cached pages as archives for research into the history of the Web.

The information storage safe harbor is not really applicable to search engines; it shields a service provider from liability where the service provider stores information at the direction of a user. Search store information provided by Web site operators, but on their own initiative rather than at the direction of the operators; this storage is more likely to fall within the system caching safe harbor or the information location tools safe harbor. Search engines may, however, store information provided by Web site operators for use in targeted advertising or paid search results, and conceivably this information could contain infringing content. The information storage safe harbor would then be available to the search engine if it does not have actual or constructive knowledge that the content was infringing, and upon acquiring such knowledge acts “expeditiously to remove, or disable access to, the material” (17 U.S.C. § 512(c)). However, the information storage safe harbor is not available to a service provider that receives “a financial benefit directly attributable to the infringing activity” (17 U.S.C. § 512(c)(1)(B)). The fact that the search engine is paid for the advertising is probably not a benefit attributable to the infringing content in the advertisement. The search engine’s advertising revenue will be the same whether the advertisement does or does not include infringing content, although where the advertisement is for something that is inherently infringing, such as (arguably) an online music file-sharing service, the ad revenue would be directly attributable to the infringement, possibly making the search engine ineligible for this safe harbor. The question seems unlikely to arise, however, as this safe harbor is designed to protect services hosting Web pages from liability for infringement posted on those pages; as with the previous two safe harbors, it is a poor fit when applied to search engines.

The information location tools safe harbor is designed with search engines in mind, although search engine operations have evolved and changed since the passage of the DMCA in 1998. The information location tools safe harbor shields a service provider from liability for referring or linking users to sites containing infringing material through tools such as directories, indexes, references, pointers, or hyperlinks. In order to be eligible for the information location tools safe harbor, the search engine or other service provider must not have actual or constructive knowledge that the located content is infringing; the “knowledge” standard here is the same as that for the information storage safe harbor. Again, like the information storage safe harbor, the information location tools safe harbor requires that the search engine must “not receive a financial benefit directly attributable to the infringing activity, in a case in which the service provider has the right and ability to control such activity” (17 U.S.C. § 512(d)(2)). This standard is drawn from the rule established by the federal courts for determining vicarious copyright infringement (see Fonovisa, 76 F.3d 259). The information location tools safe harbor includes the same notice-and-takedown procedure as the information storage safe harbor.

This notice-and-takedown procedure has given rise to controversy. The importance of search engines cannot be overestimated; a site that is overlooked by search engines will receive little or no traffic. In order to guarantee themselves the protection of the information location tools safe harbor, search engines will take down any potentially infringing links as soon as they are informed of them. This gives actual and purported copyright holders tremendous power over Web use patterns. No court order is required; the person claiming to own certain content has merely to notify the search engine of the alleged infringement, and the site will be blocked without first giving the Web site operator a chance to argue that there is no infringement. There are several reasons why a use might not be infringing: perhaps because the material is in the public domain or because the use is a fair use, or because the person claiming to be the content owner is not, in fact, the owner and the true owner has authorized the use. The operator may make these arguments after the fact, but search engines are safer ignoring such arguments until they have been resolved by a court; the requirements of the information location tools safe harbor encourage the search engine to err on the side of removing links, even at the cost of harm to the Web traffic (and thus the revenue) of the allegedly infringing Web site.

By removing the links, the search engine cuts off a sizable portion of the Web traffic that would otherwise have reached those sites. It is easy to imagine how such a procedure might be abused, especially by businesses seeking to block traffic to the Web sites of their competitors. To guard against such abuses, the DMCA establishes strict requirements for the notice that must be given to the service provider by the copyright owner:

(3) Elements of notification.—

(A) To be effective under this subsection, a notification of claimed infringement must be a written communication provided to the designated agent of a service provider that includes substantially the following:

  (i) A physical or electronic signature of a person authorized to act on behalf of the owner of an exclusive right that is allegedly infringed.

 (ii) Identification of the copyrighted work claimed to have been infringed, or, if multiple copyrighted works at a single online site are covered by a single notification, a representative list of such works at that site.

(iii) Identification of the material that is claimed to be infringing or to be the subject of infringing activity and that is to be removed or access to which is to be disabled, and information reasonably sufficient to permit the service provider to locate the material.

 (iv) Information reasonably sufficient to permit the service provider to contact the complaining party, such as an address, telephone number, and, if available, an electronic mail address at which the complaining party may be contacted.

  (v) A statement that the complaining party has a good faith belief that use of the material in the manner complained of is not authorized by the copyright owner, its agent, or the law.

 (vi) A statement that the information in the notification is accurate, and under penalty of perjury, that the complaining party is authorized to act on behalf of the owner of an exclusive right that is allegedly infringed (17 U.S.C. § 512(c)(3)(A)).

All of the information required could be faked, of course; anyone could send a letter (or, with an electronic signature, an email) claiming to be the owner of copyrighted material on a particular site. Although the person sending the letter must provide contact information, no duty is imposed upon the service provider to use that contact information. A business could thus send a letter complaining of copyrighted content on its rival’s site, claiming to be—and giving accurate contact information for—the true owner of the copyrighted material. Even though the copyright owner had actually authorized the use of its material on the site, the search engine might simply block the site without contacting the copyright owner. The business sending the letter would have violated requirements (v) and (vi), but would be unlikely to be caught; the rival’s business would suffer harm until the confusion could be straightened out—and it would be the rival who would bear the cost in time and money of straightening out that confusion.

The alleged infringer may contest the allegation of infringement by submitting a counter-notification. Within 10 to 14 business days following the filing of a proper counter-notification the service provider must restore access, unless it first receives notice that the person who submitted the original notification has filed a court action against the user. A great deal of harm can be done to a business in 10 to 14 business days, however, and the counter-notification procedure has been criticized as inadequate (Yen 2000). The search engine is given every incentive to place a higher value on the interests of the copyright holder than on any harm to the business of the alleged infringer: the DMCA insulates the search engine from liability to the Web site owner for blocking the site, so there is no risk to the search engine in doing so (17 U.S.C. § 512(g)(1)). At the same time, the DMCA shields the search engine from all monetary liability and most injunctive relief, other than orders “restraining the service provider from providing access to infringing material or activity residing at a particular online site on the provider’s system or network” (17 U.S.C. § 512(j)).

Search engines as copyright infringers

In addition to being sued as contributory or vicarious infringers for linking to allegedly copyright-infringing content, search engines may be sued as direct infringers. A search engine results page typically includes short text excerpts, and sometimes thumbnail images copied from the site on which that result is found. These are likely to increase traffic to the found sites; nonetheless, some content owners have taken exception to the practice. In 2007, the U.S. Court of Appeals for the Ninth Circuit held that the use of such thumbnail images by Google was fair use within the meaning of 17 U.S.C. § 107 (Perfect 10, 508 F.3d at 1168) and that in-line linking to the picture on the found site, even when appearing with a frame on the Google results page, was not a display of the original full-resolution picture (Perfect 10, 508 F.3d at 1161). Nor was it a copy, as no copying of the image took place on Google’s servers; the user’s computer was directed to the found site.

Search engines and trademark law

The DMCA provides search engines with a safe harbor from liability for copyright infringement; it provides no similar safe harbor from liability for trademark infringement. Section 230(c)(1) of the Communications Decency Act provides that “no provider or user of an interactive computer service shall be treated as the publisher or speaker of any information provided by another information content provider.” At one time, this was thought to provide protection for service providers from liability for some trademark infringement along lines similar to those of the DMCA copyright infringement safe harbors. However, § 230(e)(2) of the same statute provides that “nothing in this section shall be construed to limit or expand any law pertaining to intellectual property.” Sections 230(c)(1) and 230(e)(2) survived the Supreme Court’s decision in Reno v. American Civil Liberties Union, which struck down the Internet censorship portions of the Communications Decency Act as unconstitutional. The federal district court for the Southern District of New York has since explained that the effect of § 230(e)(2) is that § 230(c)(1) conveys no immunity upon service providers from liability for trademark infringement; in other words, there is no safe harbor (Gucci, 135 F. Supp. 2d 409).

Search engines may be involved in trademark infringement and dilution in many ways. They may reproduce trademark infringing text or domain names on search result pages; search engines that show images as well as text may reproduce trademark logos exactly, and the trademarks may be used in a way that infringes upon or dilutes the mark holder’s interest. And, as with copyright, search engines may provide links to sites with infringing content.

Abuse of search engines

Search engine results are crucial to the success of an online business; the higher a company shows up on a list of search results, the more hits it will receive. As a result, companies use a variety of tricks to increase search engine rank. Some of these tricks are simply sensible Web design, such as prominently displaying the terms most descriptive of the company’s business on the Web site’s front page, and publicizing the business to encourage other businesses to link to the Web site. Other tricks raise ethical questions: the use of terms, often repeated far more times than necessary simply for identification purposes, in metatags; paid placement; the use of competitors’ trademarks as a part of either of these techniques; and link farming. Unethical use of these techniques to trick search engines into ranking a site higher than it ordinarily would is called search engine spamming.

Metatags are text inserted into an HTML document so as to be invisible when the document is viewed as a Web page. To view the metatags, a user can select “View/Source” from the browser menu. Metatags, if any, will appear near the top of the HTML document, just after the title; they typically take the form <META name = “category” content = “thing(s) in category”>. While there are legitimate uses for metatags, such as to identify the Web page author and copyright holder, they have acquired notoriety for their abuse in search engine spamming. While metatags are invisible to the viewer, they are visible to search engines; thus, a term could be repeated thousands of times in metatags, potentially raising the page’s rank, without changing the visible appearance of the page.

This use of metatags is not in itself illegal; however, a business may go too far if it seeks to gain an advantage over its competitors by using the competitor’s name as a meta tag, so that searches for one Web site will detect the other as well. This use of a competitor’s name may be unlawful trademark infringement, subject to such exceptions as fair use: for example, it would probably not be infringement for an independent bicycle dealer, Citywide Bicycles, to use the trademarked name “Trek” as a meta tag, if Citywide sells or repairs Trek bicycles; this may fall within the nominative fair use exception to trademark infringement. If, however, Citywide’s chief competitor is Urban Cycles, it would be trademark infringement for Citywide to use the name Urban Cycles as a meta tag; the tag is not being used to identify Urban Cycles, but to attempt to divert business from its Web site (see Playboy v. Welles, 279 F.3d at 801). Use of metatags may be nominative fair use so long as three conditions are met:

First, the product or service in question must be one not readily identifiable without use of the trademark; second, only so much of the mark or marks may be used as is reasonably necessary to identify the product or service; and third, the user must do nothing that would, in conjunction with the mark, suggest sponsorship or endorsement by the trademark holder (279 F.3d at 801).

While technological advances have rendered meta tag spamming no longer worth the effort, the same reasoning applies to the use of protected marks in more sophisticated search engine spamming techniques.

Paid placement is one way in which companies may obtain higher search result rankings for their Web sites: the company pays a sum of money to a search engine to move its site up the rankings when a particular search term is entered. The use of trademarks as search terms raises concerns similar to those raised by the use of metatags: if Citywide Bicycles pays to be ranked at or near the top of the list when the term “Trek” is entered, there is probably no problem; the use of the name “Trek” is nominative fair use. Applying the same factors applied to metatags yields the same result: Citywide cannot identify its business (selling and repairing Trek bicycles) without using the trademark; it is using no more of the mark than necessary to identify its service; and it is not using the mark in a way that suggests the sponsorship or endorsement of the bicycle maker, Trek.

However, if Citywide pays to be ranked first when the name of its competitor, Urban Bicycles, is entered as a search term, a problem arises. This use may well constitute trademark infringement, although the question is a new enough one that it has not yet been resolved by the courts (Padawer 2003). While Google does not actively screen for trademarks used as advertising keywords (AdWords), it will act on complaints by the mark owners (Google AdWords Trademark Policy). A search engine following such a policy (analogous to the notice and takedown policy set by statute for copyright violations) is unlikely to be liable for misuse of trademark terms by others. In 2004, the Ninth Circuit Court of Appeals reversed a lower court ruling in favor of Playboy Enterprises against Netscape and Excite, which had allowed advertisers of pornographic content to link their advertisements to search results for the trademarked terms “Playboy” and “Playmate” (Playboy v. Netscape, 354 F.3d 1020). The liability of the party initiating the possibly infringing use may be, as one author observes, a bit more “difficult to predict because of conflicting opinions among jurisdictions” as to fair use and likelihood of confusion in such cases (Winkler 2012, at 1005). Results in other countries may differ as well: in a 2005 French judgment, Google was found to have infringed trademarks by allowing trademarked terms to be used as advertising keywords. The decision was upheld on appeal in 2006. Google then appealed to France’s highest court, the Court of Cassation, which in turn referred the question to the Court of Justice of the European Union. The Court of Justice held that Google, in allowing others to use trademarked terms as AdWords, was not “using” the terms within the meaning of EU copyright law, and was in any event protected from liability by provisions equivalent to the safe harbors for Internet service providers under U.S. law (Google France v Louis Vuitton, 2008). Similar suits have been filed against Google by trademark holders in Germany and Italy, but are likely to run into the same barrier.

Link farming

A link farm is “a network of pages on one or more Web sites, heavily cross-linked with each other, with the sole intention of improving the search engine ranking of those pages and sites” (Perkins 2004). Link farming is aimed at Google and related search engines that rank pages by the number and type of links to those pages from other pages, just as meta tag spamming was aimed at other search engines. Link farm spamming is considered unethical (Perkins 2004) but is not illegal; the best cure is probably not legislation but, as with metatags, advances in search engine technology that diminish the usefulness of the tactic.

Internet search engines and censorship

Search engines risk running afoul not only of copyright and trademark laws, but of censorship laws. Search engines, like all Internet resources, are accessible from nearly all of the world, and content that is legal, even accepted as mainstream, in one country may be completely illegal in another. The French Yahoo! case raised awareness of the fact that the actions of an Internet service provider in its home country may expose it to liability in another. In the aftermath of the Yahoo! case, Google has removed listings for Nazi and racist hate sites from the French and German versions of its search engines, www.google.fr and www.google.de. The power of search engines over access to information on the Internet poses the problem of private censorship; while Google blocked sites from its French and German search engines in response to complaints from the French and German governments, it has also blocked search-engine access to sites critical of the Church of Scientology at the request of the Church (McCullagh 2002). Search engines with a political agenda might even decide to remove sites from search results on their own initiative, rather than out of fear of litigation by some aggrieved outside party. (A search for “Google’s liberal bias” turns up an apparently inexhaustible vein of semiliterate paranoia.) Self-censorship of this sort poses no First Amendment concerns, as search engine operators are private companies and not state actors. However, some users’ rights advocates have begun to demand transparency in the process: search engines, they say, may remove sites from search databases if they wish, but they should publish information about which sites have been removed (McCullagh 2002). Compromise steps include Google’s promise to report legal threats (which tend to be accompanied by demands to remove or block search access to certain sites) to the free-speech Web site www.chillingeffects.org.


• Communications Decency Act, 47 U.S.C. § 230

• Digital Millennium Copyright Act, 17 U.S.C. § 512

European Union Directives

• First Council Directive 89/104/EEC of 21 December 1988 to approximate the laws of the Member States relating to trade marks, 1989 O.J. (L 40) 1

• Article 9(1) of Council Regulation (EC) No 40/94 of 20 December 1993 on the Community trade mark, 1994 O.J. (L11) 1

• Article 14 of Directive 2000/31/EC of the European Parliament and of the Council of 8 June 2000 on certain legal aspects of information society services, in particular electronic commerce, 2000 O.J. (L 178) 1


Supreme Court

Reno v. American Civil Liberties Union, 521 U.S. 844 (1997)

Federal Appellate Courts

1–800 Contacts, Inc. v. WhenU.com, Inc., 414 F.3d 400 (2d Cir. 2005)

Fonovisa, Inc. v. Cherry Auction, Inc., 76 F.3d 259 (9th Cir. 1996)

Kelly v. Arriba Soft Corp., 336 F.3d 811 (9th Cir. 2003)

Name.Space, Inc. v. Network Solutions, Inc., 202 F.3d 573 (2d Cir. 2000)

Perfect 10, Inc. v. Amazon.com, Inc., 508 F.3d 1146 (9th Cir. 2007)

Playboy Enterprises Inc. v. Netscape Communications Corp., 354 F.3d 1020 (9th Cir. 2004)

Playboy Enterprises, Inc. v. Welles, 279 F.3d 796 (9th Cir. 2002)

Yahoo! Inc. v. La LigueContre Le RacismeetL’Antisemitisme, 433 F.3d 1199 (9th Cir. 2006), reversed and remanded, Yahoo! Inc. v. La LigueContre Le Racisme et L’Antisemitisme, 379 F.3d 1120 (9th Cir. 2004), rehearing, 433 F.3d 1199 (9th Cir. 2006), certiorari denied, 547 U.S. 1163 (2006)

Federal Trial Court

Gucci America, Inc. v. Hall & Associates, 135 F. Supp. 2d 409 (S.D.N.Y. 2001)

European Union

• Case C-236/08, Google France v. Louis Vuitton, Judgment of Court (Grand Chamber) (March 23, 2010), available at http://curia.europa.eu/juris/liste.jsf?language=en&num=C-236/08 (visited February 18, 2013)

See also Censorship; Communications Decency Act; Copyright; Copyright Infringement; Data Mining; Digital Millennium Copyright Act, Title II; File-Sharing; French Yahoo! Case; Metatags; Trademark Dilution; Trademark Infringement

Sources and Further Reading

“An Ad-Block Shock: France v Google,” The Economist (January 12, 2013), available at http://www.economist.com/news/business/21569414-xavier-niel-playing-rough-internet-giant-france-v-google (visited February 18, 2013)

“AdWords Trademark Policy,” Google, available at http://support.google.com/adwordspolicy/bin/answer.py?hl=en&answer=6118 (visited December 20, 2012)

Jesse Drucker, “Google Shifts $10b into Bermuda Subsidiary Tax Bill,” Royal Gazette Online (December 11, 2012, 8:00 AM), available at http://www.royalgazette.com/article/20121211/BUSINESS/712109923 (visited February 18, 2013)

“Google and Antitrust: Transatlantic Tussles,” The Economist (December 22, 2012), at 105, available at http://www.economist.com/news/business/21568757-investigations-google-approach-climax-america-and-europe-transatlantic-tussles (visited February 18, 2013)

“Google’s Copyright Complaints Flag Up Piracy of Microsoft,” BBC News (May 25, 2012), available at http://www.bbc.co.uk/news/technology-18204411 (visited February 8, 2013)

“Google in Africa: It’s a Hit,” The Economist (May 12, 2012), at 57, available at http://www.economist.com/node/21554566 (visited February 18, 2013)

“Google Threatens French Media Ban over Proposed Law,” BBC News (October 18, 2012), available at http://www.bbc.co.uk/news/technology-19996351 (visited February 18, 2013)

Jessica Leber, “EmTech: French Minister Says Google Faces Critical Questions over Copyright,” MIT Technology Review (October 25, 2012), available at http://www.technologyreview.com/view/506296/emtech-french-minister-says-google-faces-critical-questions-over-copyright/ (visited February 18, 2013)

Declan McCullagh, “Google Excluding Controversial Sites,” CNET News.com (October 23, 2002), available at http://news.cnet.com/2100–1023–963132.html (visited February 18, 2013)

Heidi S. Padawer, “Google This: Search Engine Results Weave a Web for Trademark Infringement Actions on the Internet,” 81 Washington University Law Quarterly 1099 (2003)

Alan Perkins, “The Classification of Search Engine Spam,” Search Mechanics, available at http://www.ebrandmanagement.com/whitepapers/spam-classification/ (visited October 28, 2004)

“Search Tools Listing: Open Source Search Engines,” Search Tools, available at http://www.searchtools.com/tools/tools-opensource.html (visited February 18, 2013)

Craig W. Walker, “Application of the DMCA Safe Harbor Provisions to Search Engines,” 9 Virginia Journal of Law & Technology 1 (2004)

Katelynn Winkler, “Functionality Failure: How Rosetta Stone v. Google Misapplies the Trademark Functionality Doctrine to AdWords,” 44 Arizona State Law Journal 995 (2012)

Alfred C. Yen, “Internet Service Provider Liability for Subscriber Copyright Infringement, Enterprise Liability, and the First Amendment,” 88 Georgetown Law Journal 1833 (2000)


See Encryption


See Consumer Broadband and Digital Television Promotion Act


A semiconductor manufacturing mask work is a design pattern used in the manufacture of computer chips. These chips are manufactured in a three-stage process. The first stage is the manufacture of the blank wafers from which the chips will be etched; the wafers are usually made of silicon, gallium arsenide, or indium phosphide, and are made at a plant solely dedicated to wafer manufacturing; the second stage—the transformation of the blank wafer into a chip–generally takes place elsewhere. The third stage is the assembly and packaging; the etched wafers are cut apart and mounted in a ceramic or metal package, and leads are soldered on (Lee & Elliot 2006).

The second stage of the process, the transformation of blank wafers into computer chips, is the most technically complex. The surface of the wafer is first oxidized by being exposed to oxygen at a high temperature. Much of the rest of the manufacturing process is devoted to selectively removing portions of this oxidized layer. This is done by first applying a coating of photoresist, a light-sensitive compound, over the oxidized layer. Portions of the photoresist are then exposed to light through a mask (Lee & Elliot n.d.). The mask is the informational portion of the manufacturing process; it determines which parts of the photoresist shall be exposed, in the same way, although not by the identical process, that a laser printer, photocopier, or photographic negative determines what part of a print drum or photographic paper shall be exposed. In essence, an image of the mask is being recorded on the photoresist-coated surface of the wafer.

Exposure to light causes a portion of the photoresist to either harden or soften. (Photoresist may be either positive or negative: positive photoresist softens when exposed to light, while negative photoresist hardens.) A photoresist-developing chemical is then used to strip away the soft portion of the photoresist; the hard portion remains. Where the photoresist has been removed, the silicon dioxide layer is exposed. This layer is then etched by exposure to acid or a reactive gas plasma. The channels left by the etching are then “doped”—conducting and semiconducting materials are deposited in them by one or more of several processes. A new layer of silicon is then deposited on the surface of the wafer, and the process begins again for as many layers as necessary. Finally, the wafers are polished and enter the third stage of production (Lee & Elliot n.d.).

The semiconductor manufacturing mask work is crucial to the process; the design of the mask is a long and complex process. Once designed, however, it can be imitated by reverse engineering. Conceivably a mask work could be made the subject of copyright or patent, but instead Congress has chosen to create a new category of intellectual property, giving the mask work copyright-like protection. The Semiconductor Chip Protection Act of 1984 provides for the registration of semiconductor mask work designs with the U.S. Copyright Office (17 U.S.C. § 908; 37 C.F.R. § 211.4). The registrant must then alert the world to the claim of an intellectual property right in the mask work by placing the words “mask work,” the symbol “M,” or the letter “M” in a circle in a visible location on the exterior surface of the finished chip or its receptacle (17 U.S.C. § 909; 37 C.F.R. § 211.6). The registrant will enjoy the exclusive right to reproduce the mask work and to import or distribute a semiconductor chip embodying the mask work, as well as the exclusive right to authorize, induce, or knowingly cause any other person to do either of these things (17 U.S.C. § 905). This exclusive right will last for 10 years from the date of registration or from “the date on which the mask work is first commercially exploited anywhere in the world, whichever occurs first” (17 U.S.C. § 904). The Act also includes a right of first sale (17 U.S.C. § 906(b)) and an exception to the exclusive rights for reverse engineering (17 U.S.C. § 906(a)). Innocent infringers are also protected (17 U.S.C. § 907).

Semiconductor manufacturing masks are not like other protected works such as MP3s; they are not subject to widespread piracy and infringement. Anyone can copy and use the MP3, but without a chip-manufacturing factory, the mask is useless. Chip factories are relatively rare, and setting up a new factory may require an investment of hundreds of millions of dollars. As a result, infringements of registered mask works are rare. The Semiconductor Chip Protection Act has given rise to relatively little litigation, resulting in only a handful of reported cases, only one of which—Brooktree v. AMD—deals with the substantive application of the Act. The others include Anadigics v. Raytheon and Atari v. Nintendo. The reported opinion in Anadigics dealt with procedural issues, however (903 F. Supp. 615), while the court in Atari concluded that the Act did not apply (975 F.2d at 842 n. 5).

Brooktree v. Advanced Micro Devices involved the reverse engineering exception. The defendant, Advanced Micro Devices (AMD), claimed to have reverse-engineered the chip, resulting in a different chip that performed the same function. The court agreed that such reverse engineering would be permitted under the Act. However, at trial, a jury had found that AMD’s chip design did infringe upon Brooktree’s registered mask work; the appellate court held that a reasonable jury could have reached this conclusion, and declined to reverse the trial court’s decision. The appellate court pointed out that, as a witness had testified to Congress in hearings on the Act,

Whenever there is a true case of reverse engineering, the second firm will have prepared a great deal of paper—logic and circuit diagrams, trial layouts, computer simulations of the chip, and the like; it will also have invested thousands of hours of work. All of these can be documented by reference to the firm’s ordinary business records. A pirate has no such papers, for the pirate does none of this work. Therefore, whether there has been a true reverse engineering job or just a job of copying can be shown by looking at the defendant’s records. The paper trail of a chip tells a discerning observer whether the chip is a copy or embodies the effort of reverse engineering (Brooktree, 977 F.2d at 1566).