by Mol­ly Bal­ti­more 1

Tech­no­log­i­cal advance­ment over the past few decades has allowed the pub­lic to access a vast wealth of knowl­edge and infor­ma­tion at the click of a but­ton. We no longer have to trav­el to a library or con­sult a pro­fes­sion­al to fact-check a friend’s argu­ment or become a qua­si-expert in a sub­ject area. Con­sid­er­ing the depth and com­plex­i­ty of the infor­ma­tion avail­able, nav­i­ga­tion­al tools such as online data­bas­es have become indis­pens­able to Inter­net users. How­ev­er, the pres­ence of copy­right­ed mate­r­i­al on the World Wide Web presents dif­fi­cult legal and pol­i­cy ques­tions. If we think copy­right­ed infor­ma­tion should be freely avail­able to the pub­lic via online data­bas­es, where do we draw the line? This arti­cle will argue that cer­tain online text-search­able data­bas­es can con­sti­tute fair use even if they pro­vide users copy­right­ed infor­ma­tion. Specif­i­cal­ly, search engines that serve a trans­for­ma­tive pur­pose and lim­it the amount of pro­tect­ed mate­r­i­al avail­able to the user—such as the Google Books database—should con­sti­tute fair use.

Copy­right exists in order “[t]o pro­mote the Progress of Sci­ence and use­ful Arts.” 2 Afford­ing cre­ators lim­it­ed exclu­sive rights over their work cre­ates an incen­tive for artists to cre­ate with­out fear of exploita­tion. How­ev­er, there are lim­its to this pro­tec­tion. The fair use defense allows sec­ondary users to copy or repro­duce oth­er authors’ works with­out being liable for copy­right infringe­ment in cer­tain instances, usu­al­ly where the sec­ondary work serves as com­men­tary on or a trans­for­ma­tive appro­pri­a­tion of the orig­i­nal. 3

Rel­a­tive­ly recent tech­no­log­i­cal inno­va­tion, par­tic­u­lar­ly the growth of the Inter­net, is requir­ing courts to con­front nov­el ques­tions of fair use. Specif­i­cal­ly, the cre­ation and expan­sion of online text-search­able data­bas­es con­tain­ing copy­right­ed infor­ma­tion presents a unique legal quandary that has yet to be addressed by the Supreme Court. 4 For exam­ple, the Google Books data­base allows users to read ver­ba­tim excerpts from copy­right­ed works. 5 This tool and oth­er sim­i­lar online cat­a­logues serve a use­ful func­tion by allow­ing users to pre­view a vari­ety of mate­ri­als to aid in tar­get­ed research efforts. How­ev­er, they also flirt with the bound­aries of fair use by deliv­er­ing a sig­nif­i­cant amount of pro­tect­ed mate­r­i­al to users at no cost and with­out a license from the hold­ers of copy­right. 6 As one schol­ar not­ed, these online cat­e­go­riz­ers have become a “vir­tu­al neces­si­ty in a com­plex maze of infor­ma­tion,” and courts must find a way to accom­mo­date their use while pro­tect­ing expres­sive copy­right­ed lan­guage. 7


To fur­ther the objec­tive of copy­right, courts have allowed fair uses of copy­right­ed work for near­ly three cen­turies. The jus­ti­fi­ca­tion for this doc­trine is that, while it is nec­es­sary to pro­tect cre­ators’ rights and inter­ests, “the ulti­mate, pri­ma­ry intend­ed ben­e­fi­cia­ry is the pub­lic, whose access to knowl­edge copy­right seeks to advance.” 8 Deter­min­ing fair use requires a fact-dri­ven analy­sis that del­i­cate­ly bal­ances the civic goals of copy­right with the inter­ests of authors. The fair use doc­trine was cod­i­fied in sec­tion 107 of the Copy­right Act of 1976, which “allowed the fair use of a copy­right­ed work . . . for pur­pos­es such as crit­i­cism, com­ment, news report­ing, teach­ing . . . schol­ar­ship, or research.” 9 The statute laid out the fol­low­ing four-fac­tored analy­sis to deter­mine whether a par­tic­u­lar instance of copy­ing is con­sid­ered a fair use:

  1. the pur­pose and char­ac­ter of the use, includ­ing whether such use is of a com­mer­cial nature or is for non­prof­it edu­ca­tion­al purposes;
  2. the nature of the copy­right­ed work;
  3. the amount and sub­stan­tial­i­ty of the por­tion used in rela­tion to the copy­right­ed work as a whole; and
  4. the effect of the use upon the poten­tial mar­ket for or val­ue of the copy­right­ed work. 10

The Supreme Court has rec­og­nized that in enact­ing sec­tion 107, Con­gress intend­ed to cod­i­fy the com­mon law doc­trine of fair use, and “not to change, nar­row, or enlarge it in any way.” 11 Both the statute and the doc­trine of fair use require a case-by-case analy­sis in which each of the four statu­to­ry fac­tors are “explored, and the results weighed togeth­er, in light of the pur­pos­es of copy­right.” 12 How­ev­er, the Supreme Court has stressed that some of the fac­tors car­ry more sig­nif­i­cance than oth­ers. Specif­i­cal­ly, the first fac­tor, which assess­es the “pur­pose and char­ac­ter of the use,” 13 is impor­tant. “[T]he more trans­for­ma­tive the new work, the less will be the sig­nif­i­cance of oth­er fac­tors, like com­mer­cial­ism, that may weigh against a find­ing of fair use.” 14 Addi­tion­al­ly, the fourth fac­tor, which con­sid­ers whether the sec­ondary use has caused harm to the mar­ket for or val­ue of the copy­right­ed work, “is undoubt­ed­ly the sin­gle most impor­tant ele­ment of fair use.” 15

The large amount of flex­i­bil­i­ty embed­ded in this four-fac­tor analy­sis presents advan­tages and dis­ad­van­tages. Courts are free to make deter­mi­na­tions of fair use on a case-by-case basis, and there “is no dis­po­si­tion to freeze the doc­trine in the statute, espe­cial­ly dur­ing a peri­od of rapid tech­no­log­i­cal change.” 16 At the same time, this lat­i­tude cre­ates uncer­tain­ty, and the statute may be applied dif­fer­ent­ly to sim­i­lar fac­tu­al sce­nar­ios, based on a judge’s ide­ol­o­gy. 17


With respect to online data­bas­es, the fair use defense should be read more expan­sive­ly. These tools are unique in that they exist to advance the very pur­pose of copyright—the typ­i­cal use of Google Books and sim­i­lar search engines is edu­ca­tion­al and schol­ar­ly in nature. 18 While the Supreme Court has nev­er decid­ed a case con­cern­ing the appli­ca­tion of fair use doc­trine to an online data­base, the Sec­ond, Fourth, and Ninth Cir­cuits have upheld them as fair use.

In Per­fect 10, Inc. v., Inc., an adult enter­tain­ment mag­a­zine sued the cre­ators of an online search engine for copy­right infringe­ment, claim­ing that the search engine pro­vid­ed thumb­nail copies of pro­tect­ed images. 19 Regard­ing the first fac­tor of the fair use test, the Ninth Cir­cuit held that the search engine con­sti­tut­ed a trans­for­ma­tive use, as “even mak­ing an exact copy of a work may be trans­for­ma­tive so long as the copy serves a dif­fer­ent func­tion than the orig­i­nal.” 20 The court rea­soned that the search engine served the pur­pose of allow­ing users to iden­ti­fy pho­tos of inter­est to direct them to their orig­i­nat­ing web­page, where­as the orig­i­nal pho­tos them­selves served an enter­tain­ment or aes­thet­ic pur­pose. 21 Like­wise, the Fourth Cir­cuit held that a sec­ondary use of stu­dent papers in an online pla­gia­rism detec­tion data­base was trans­for­ma­tive and con­sti­tut­ed fair use. 22

The Sec­ond Cir­cuit has recent­ly decid­ed two cas­es con­cern­ing online text-search­able data­bas­es. In Authors Guild, Inc. v. HathiTrust, indi­vid­ual authors and orga­ni­za­tions of authors claimed that an online dig­i­tal library infringed their copy­rights by scan­ning the full text of books into a data­base. 23 The court held the use to be trans­for­ma­tive, as the data­base pro­vid­ed infor­ma­tion about the books. 24 How­ev­er, impor­tant to the hold­ing in HathiTrust was the fact that the data­base did not reveal any text from the books, which was the sit­u­a­tion in Authors Guild v. Google, Inc.

In Google, the Sec­ond Cir­cuit rea­soned that the text snip­pets revealed by the data­base were nec­es­sary to give the searcher enough—but not too much—context to deter­mine whether the copy­right­ed work “falls with­in the scope of [the user’s] inter­est” and, thus, whether he should pur­chase it. 25 Fur­ther­more, the data­base imposed sig­nif­i­cant lim­i­ta­tions on the amount of infor­ma­tion made avail­able to the user. The data­base only dis­played snip­pets that were approx­i­mate­ly one-eighth of a page in length and blocked one snip­pet per page as well as one com­plete page for every ten pages. 26 Regard­ing the fourth and “most impor­tant” fac­tor, the Google court acknowl­edged that “the more the copy­ing is done to achieve a pur­pose that dif­fers from the pur­pose of the orig­i­nal, the less like­ly it is that the copy will serve as a sat­is­fac­to­ry [mar­ket] sub­sti­tute for the orig­i­nal.” 27 Thus, it is less like­ly that a trans­for­ma­tive use will result in a direct eco­nom­ic harm to the orig­i­nal. Instead of ana­lyz­ing the fourth fac­tor on the extreme­ly broad “poten­tial” eco­nom­ic harm to the copy­right hold­er, courts should require evi­dence of real harm. 28

Fair use has also been impli­cat­ed when ref­er­ence guides uti­lize por­tions of copy­right­ed mate­r­i­al. For exam­ple, in Warn­er Bros. Entm’t Inc. v. RDR Books, an enthu­si­as­tic fan of the Har­ry Pot­ter book series cre­at­ed an online ency­clo­pe­dia and print “Lex­i­con” that served as ref­er­ence guides, allow­ing read­ers of the series to eas­i­ly nav­i­gate par­tic­u­lar por­tions of the books. 29 While the court found that the sec­ondary work was trans­for­ma­tive in nature, the degree of ver­ba­tim copy­ing of the orig­i­nal works was such that it weighed against a find­ing of fair use. How­ev­er, schol­ars have observed that the ques­tion remains whether the court would have found the Lex­i­con to be fair use had it exist­ed pure­ly in its elec­tron­ic form. 30


As these cas­es sug­gest, online data­bas­es that con­vey copy­right­ed infor­ma­tion can do so in a trans­for­ma­tive man­ner with­out caus­ing real eco­nom­ic harm. Such data­bas­es pro­vide infor­ma­tion about works of author­ship. Rather than serv­ing the same enter­tain­ment or aes­thet­ic pur­pose as the under­ly­ing mate­r­i­al, they oper­ate in a mar­ket dis­tinct from that for the orig­i­nal work. Giv­en that every fed­er­al court of appeals to rule on the issue has upheld these data­bas­es as fair use, the Supreme Court should fol­low suit if giv­en the chance.


  1. The fol­low­ing arti­cle reflects my expe­ri­ence in the 2016 Car­do­zo BMI Moot Court Com­pe­ti­tion. The prob­lem involved the cre­ator of an online text-search­able data­base based on a fic­ti­tious, pop­u­lar book and film series called Char­lie Trot­ter. The copy­right hold­ers of the book and film series filed suit alleg­ing copy­right infringe­ment, as the data­base cre­ator copied the books into the data­base and revealed text snip­pets to users at no cost. The ques­tion pre­sent­ed was whether the alleged infringer was enti­tled to the fair use defense under sec­tion 107 of the Copy­right Act.
  2. U.S. Con­st., art. I, § 8, cl. 8.
  3. See 17 U.S.C. § 107; Camp­bell v. Acuff-Rose Music, Inc., 510 U.S. 569 (1994); Harp­er & Row, Pub­lish­ers, Inc. v. Nation Enters., 471 U.S. 539 (1985).
  4. See, e.g., Authors Guild v. Google, Inc., 804 F.3d 202 (2d Cir. 2015), cert. denied, 136 S. Ct. 1658 (2016); Authors Guild, Inc. v. HathiTrust, 755 F.3d 87 (2d Cir. 2014).
  5. See Google, 804 F.3d at 202.
  6. See HathiTrust, 755 F.3d at 98; Google, 804 F.3d at 221; see also Per­fect 10, Inc. v., Inc., 508 F.3d 1146, 1165 (9th Cir. 2007); Bill Gra­ham Archives v. Dor­ling Kinder­s­ley Ltd., 448 F.3d 605, 612 (2d Cir. 2006).
  7. Andrew Sar­rol, The Copy­right Impli­ca­tions of Search­able Data­bas­es: A Method­ol­o­gy for Ana­lyz­ing the Fourth Fair Use Fac­tor, 3 Seton Hall Cir­cuit Rev. 527, 529 (2007).
  8. Google, 804 F.3d at 212.
  9. 17 U.S.C. § 107.
  10. Id.
  11. Camp­bell v. Acuff-Rose Music, Inc., 510 U.S. 569, 577 (1994) (quot­ing H.R. Rep. No. 94–1476, at 66 (1976), as reprint­ed in 1976 U.S.C.C.A.N. 5659, 5680).
  12. Camp­bell, 510 U.S. at 578 (cita­tions omit­ted).
  13. 17 U.S.C. § 107(1).
  14. Camp­bell, 510 U.S. at 579.
  15. Harp­er & Row, Pub­lish­ers, Inc. v. Nation Enters., 471 U.S. 539, 566 (1985).
  16. Sony Corp. of Am. v. Uni­ver­sal City Stu­dios, Inc., 464 U.S. 417, 448 n.31 (1984).
  17. Nari Na, Test­ing the Bound­aries of Copy­right Pro­tec­tion: The Google Books Library Project and the Fair Use Doc­trine, 16 Cor­nell J.L. & Pub. Pol’y 417, 433 (2007).
  18. Han­ni­bal Travis, Build­ing Uni­ver­sal Dig­i­tal Libraries: An Agen­da for Copy­right Reform, 33 Pepp. L. Rev. 761, 822 (2006).
  19. Per­fect 10, 508 F.3d 1146.
  20. Id. at 1165.
  21. Id.
  22. See A.V. ex rel. Van­der­hye v. iPar­a­digms, LLC, 562 F.3d 630 (4th Cir. 2009).
  23. HathiTrust, 755 F.3d at 87.
  24. Id. at 97.
  25. Google, 804 F.3d at 217–18.
  26. Id. at 210.
  27. Id. at 223.
  28. Travis, supra note 18, at 814.
  29. 575 F. Supp. 2d 513 (S.D.N.Y. 2008).
  30. Shi­ra Siskind, Cross­ing the Fair Use Line: The Demise and Revival of the Har­ry Pot­ter Lex­i­con and Its Impli­ca­tions for the Fair Use Doc­trine in the Real World and on the Inter­net, 27 Car­do­zo Arts & Ent. L. J. 291, 307 (2009).