Legal aspects of text mining

Computer Law and Security Review

Volumn 30, Issue 2, 2014, Pages 153-170

  Truyens, Maarten ^a   Van Eecke, Patrick ^a,b,c  

^a UNIVERSITY OF ANTWERP   (Belgium)

^b KING'S COLLEGE LONDON   (United Kingdom)

^c QUEEN MARY UNIVERSITY OF LONDON   (United Kingdom)

Author keywords

Copyright; Data mining; Databases; Reproduction right; Text mining

Indexed keywords

COPYRIGHTS; DATABASE SYSTEMS;

CONTRACT LAWS; COPYRIGHT LAW; EXTRACT INFORMATIONS; LEGAL ASPECTS; MULTIPLE DISCIPLINES; REPRODUCTION RIGHT; TEXT DOCUMENT; TEXT MINING;

DATA MINING;

EID: 84897917103     PISSN: 02673649     EISSN: None     Source Type: Journal    
DOI: 10.1016/j.clsr.2014.01.009     Document Type: Article

References (170)

1. R. FELDMAN and J. SANGER, The text mining handbook: advanced approaches in analyzing unstructured data, New York, Cambridge University Press, 2007, 1.
2. M. HEARST, University of Maryland, Untangling Text Data Mining, Proceedings of the ACL '99, 1999.
3. G. MINER, Practical text mining and statistical analysis for nonstructured text data applications, Waltham, MA, Academic Press, 2012, 31.
4. M. KONCHADY, Text mining application programming, Boston, Mass., Charles River Media, 2006, 3.
5. P. KOEHN, Statistical machine translation: textbook, Cambridge, Univ. Pr., 2007, 5-6.
6. See for example the "AMiCA" project (www.amicaproject.be).
7. R. MOCHALES-PALAU and M.-F. MOENS, "Argumentation mining," Artificial Intelligence and Law 19, 1;
8. A. WYNER, R. MOCHALES-PALAU, M.-F. MOENS and D. MILWARD, "Approaches to Text Mining Arguments from Legal Cases," in E. FRANCESCONI, S. MONTEMAGNI, W. PETERS and D. TISCORNIA (eds.), Semantic Processing of Legal Texts, Springer Berlin Heidelberg, January 1, 2010.
9. C. VANDA BURNS, "Online legal services d a revolution that failed?" October 2007.
10. Example cited by a user submission in I. HARGREAVES, Supporting Document T to the Hargreaves report e Text Mining and Data Analytics in Call for Evidence Responses, May 2011, www.ipo.gov.uk/ipreview-doc-t.pdf, 5.
11. For linguistic corpora, see for example the Linguistic Data Consortium www.ldc.upenn.edu/.
12. J. PUSTEJOVSKY and A. STUBBS, Natural Language Annotation for Machine Learning, O'Reilly Media Inc., October 10, 2012, 8.
13. G. LEECH, Corpus annotation: Linguistic information from computer text corpora, 1997, 1-18.
14. Supervised learning typically provides higher quality output, but may have a significantly higher cost due to the amount of human effort required. Practice has shown however that, when certain assumptions can be made about the set of humanprovided examples (e.g., regarding their smooth distribution), using unlabelled data in conjunction with a small amount of labelled data can produce very decent results. Semi-supervised learning will be most useful whenever obtaining data points is cheap, but obtaining the human-provided labels costs a lot of time, effort, or money. This is the case in many application areas of machine learning (e.g., billions of web pages are readily available, but to classify them reliably, humans have to read them). See O. CHAPELLE, B. SCHöLKOPF and A. ZIEN, Semi-supervised learning, Cambridge, Mass., MIT Press, 2006, 1-11.
15. J. PUSTEJOVSKY, o.c., 15-16.
16. C.D. MANNING, P. RAGHAVAN and H. SCHÜTZE, Introduction to information retrieval, New York, Cambridge University Press, 2008, 350.
17. In fact, because decision trees are essentially a set of questions that need to be answered, it is very common for domain experts to start from a software-generated decision tree to understand the key factors, and then refine the tree to better match his or her own beliefs. See T. SEGARAN, Programming collective intelligence: building smart web 2.0 applications, Beijing; Sebastapol [Calif.], O'Reilly, 2007, 155.
18. See the announcement at googlesystem.blogspot.be/2008/05/ using-googles-n-gram-corpus.html.
19. Communication from the Commission on content in the Digital Single Market, 18 December 2012 (COM (2012) 789 final), available at goo.gl/dc64Xq.
20. For a general discussion of text mining (in particular from a legalephilosophical perspective) in the context of archiving and digitization: M. BORGHI and S. KARAPAPA, Copyright and Mass Digitization, Oxford, Oxford University Press, 2013, 45-69.
21. In addition, the Hargreaves review in the United Kingdom (available at www.ipo.gov.uk/types/hargreaves.htm) about the revision of UK intellectual property legislation also touches upon text mining, and has in fact resulted in legislative amendments (see below). A similar revision process is currently undertaken in Australia, and is currently in the stage of a public consultation process e for an analysis of text mining
22. see goo.gl/dStpVV.
23. See, inter alia, L. BYGRAVE, "Automated profiling: minding the machine: article 15 of the EC Data Protection Directive and Automated Profiling," Computer Law & Security Review 17, 1;
24. B.W. SCHERMER, "The limits of privacy in automated profiling and data mining," Computer Law & Security Review 27, 1, February 2011.
25. A similar situation applies in France. For an overview of several countries, see en.wikipedia.org/wiki/Official- text-copyright.
26. For a general overview of the history and current assessment of the concept of originality (with particular attention to the CJEU, UK and US points of view), see the PhD thesis of Eleonora Rosati: E. ROSATI, Judge-made EU Copyright Harmonisation e the case of originality, PhD thesis, European University Institute, 2012.
27. For a general comparison, see K. HARIANI and A. HARIANI, "Analyzing Originality in Copyright Law: Transcending Jurisdictional Disparity," IDEA 51, 2011.
28. In the UK, labour is essentially at the basis of copyright protection (E. ROSATI, o.c., 80.). According to the authoritative case (London Press, Ltd v University Tutorial Press, Ltd., 1911, 608), in order to benefit from copyright protection, a work must not be copied from another work: it should instead originate from the author.
29. Walter v. Lane, 1900, AC 539;
30. Express Newspapers v News (UK) Ltd., 1990, 1 W.L.R. 1320
31. Accordingly, the threshold towards copyright protection is rather low. The US Supreme Court rejected such low threshold in the seminal Feist case (Feist Publications, Inc., v Rural Telephone Service Co., 499 U.S. 340 (1991)), and ruled that copyright protection was not granted to works that do not expose any kind of creative input (such as, in the case at hand, telephone directories). According to the US Supreme Court, a minimal spark of creativity is required, so that no one may claim copyright in facts or ideas, such as telephone numbers.
32. W. GROSHEIDE, "Moral rights", in E. DERCLAYE (ed.), Research handbook on the future of EU copyright, 2009, 242.
33. Infopaq v Danske Dagblades Forening, C-5/08, 16 July 2009 ("Infopaq I");
34. Bezpe-cnostní softwarová asociace v Ministerstvo Kultury, C-393/09, 22 December 2010;
35. Premier League, joint cases C-403/08 and C-429/08, 4 October 2011;
36. Eva-Maria Painer v Standard VerlagsGmbH e.a., C-145/10, 1 December 2011;
37. Infopaq v Danske Dagblades Forening, C-302/10, 17 January 2010 ("Infopaq II");
38. Football Dataco v Yahoo!, C-604/10, 1 March 2012.
39. "That criterion of originality is satisfied when (...) its author expresses his creative ability in an original manner by making free and creative choices and thus stamps his "personal touch"" e ECJ March 1, 2012
40. Football Dataco Ltd and Others v Yahoo! UK Ltd and Others, Case C-604/10, paragraph 38.
41. "the significant labour and skill (...) cannot as such justify such a protection if they do not express any originality" e Footnote 33, 46.
42. Infopaq I, paragraph 45.
43. Football Dataco (C-604/10), paragraphs 38-40.
44. For example, football matches are subject to the rules of the game, leaving no room for creative freedom for the purposes of copyright. See the Premier League joined cases (C-403/08 and C- 429/08), paragraphs 96-99.
45. Painer, paragraphs 97-99. See also: B. MICHAUX and F. DE VISSCHER, Précis du droit d'auteur et des droits voisins, 2000, 18-19.
46. E. ROSATI, Judge-made EU Copyright Harmonisation e the case of originality, PhD thesis, European University Institute, 2012, 101.
47. T. APLIN, "United Kingdom", in B. LINDNER and T. SHAPIRO (eds.), Copyright in the Information Society. A guide to national implementation of the European Directive, 2011, 558-560.
48. E. DERCLAYE, "Infopaq International A/S v Danske Dagblades Forening (C-5/08): wonderful or worrisome?," European Intellectual Property Review 32, 5, 2010.
49. Meanwhile, and despite the CJEU decision, the Belgian Supreme Court (Cour de Cassation) issued a decision on 26 January 2012 in which the "personal touch" requirement was explicitly abolished, despite the fact that Belgian case law and legal doctrine had explicitly included this requirement for many years. Several legal commentators assumed that the Court's decision was an unfortunate mistake. See H. VANHEES, "Originaliteit in het auteursrecht" (noot onder Cass. 26 januari 2012), RW 2012-13, December 8, 2012.
50. See the official communication (with references to parliamentary documents) at goo.gl/jZzPTW.
51. For background commentary, see B. HUGENHOLTZ, "Goodbye, Geschriftenbes cherming!" Kluwer Copyright Blog, http:// kluwercopyrightblog.com/2013/03/06/goodbye-geschriftenbescherming/.
52. For a general discussion of the geschriftenbescherming, see A. BEUNEN, "Geschriftenbescherming: The Dutch Protection for Non-original Writings," A Century of Dutch Copyright Law e Auteurswet 1912-2012, DeLex, 2012.
53. M. VAN EECHOUD, "Along the Road to Uniformity e Diverse Readings of the Court of Justice Judgments on Copyright Work," JIPITEC 2012, 1, paragraph 74.
54. Infopaq I, paragraphs 37-39.
55. G. WESTKAMP, The implementation of Directive 2001/29/EC in the Member States, Part II, February 2007, goo.gl/Iea1EJ
56. Article 5(3)(a) of the Copyright Directive.
57. P.B. HUGENHOLTZ and M.R.F. SENFTLEBEN, Fair use in Europe. In search of flexibilities, November 2011, 14.
58. For an overview of the changes in the traditional dichotomy of university v. industry research, see P. VAN EECKE, J. KELLY, P. BOLGER and M. TRUYENS, Monitoring and Analysis of Technology Transfer and Intellectual Property Regimes and Their Use, Story Publishers, 2008, http://goo.gl/1YYBgX, 238.
59. I. HARGREAVES, Supporting Document T to the Hargreaves report e Text Mining and Data Analytics in Call for Evidence Responses, May 2011, www.ipo.gov.uk/ipreview-doc-t.pdf, 7-9.
60. G. WESTKAMP, The implementation of Directive 2001/29/EC in the Member States, Part II, February 2007, goo.gl/Iea1EJ, 32-35.
61. Article 5(3)(d) of the Copyright Directive.
62. For a comparison between the Dutch and the German quotation exception, see M. SENFTLEBEN, "Bridging the Differences between Copyright's Legal TraditionseThe Emerging EC Fair Use Doctrine," Journal of the Copyright Society of the USA 57, 3, 2010, 535-538.
63. Assuming that no specific contractual conditions were agreed. See UK Supreme Court April 12, 2013, Public Relations Consultants Association Limited v The Newspaper Licensing Agency Limited and others, 21.
64. Explanatory Memorandum for the Copyright Directive Proposal (COM (1997) 628), art. 5.
65. Study on the implementation and effect in Member States' laws of Directive 2001/29/EC on the harmonisation of certain aspects of copyright and related rights in the information society, February 2007, www.ivir.nl/ publications/guibault/Infosoc-report- 2007.pdf, 30.
66. Newspaper Licensing Agency Ltd and others v Meltwater Holding BV and others, 2010, EWHC 3099 (Ch);
67. EWCA, 2011, Civ 890.
68. For a general discussion of this case, see I. HARDING, "Is NLA v Meltwater the end of browsing?," Journal of Intellectual Property Law & Practice 7, 7, June 2012.
69. Infopaq I, paragraph 64.
70. Infopaq II, paragraph 32.
71. S. DEPREEUW, "De uitzondering voor 'tijdelijke technische reproductiehandelingen' na Infopaq I en II en Premier League," Auteurs & Media 2013, 2, 82.
72. According to some commentators, such permission would be granted by the mere act of knowingly uploading content onto an open web page. See Study on the implementation and effect in Member States' laws of Directive 2001/29/EC on the harmonisation of certain aspects of copyright and related rights in the information society, February 2007, www.ivir.nl/publications/ guibault/Infosoc-report-2007.pdf, 34.
73. Premier League, paragraph 171.
74. Infopaq II, paragraphs 51-53.
75. Infopaq II, paragraph 57.
76. Infopaq I, paragraph 56.
77. Football Dataco, paragraphs 27-28.
78. Football Dataco, paragraphs 47-52.
79. Article 7(1) Database Directive; British Horseracing Board case v William Hill, C-203/02, 9 November 2004, paragraphs 23-42.
80. Some additional elements (e.g., whether computer-generated databases qualify for sui generis protection) may also need to be considered, but are not within the scope of this contribution. See M. SCHELLEKENS, "A database right in search results? e An intellectual property right reconsidered in respect of computer generated databases," Computer Law & Security Review 27, 6, December 2011.
81. Fixtures Marketing Ltd v Svenska Spel AB, C-338/02, 9 November 2004, paragraph 24.
82. Fixtures Marketing, paragraphs 29-37.
83. Fixtures Marketing, paragraph 28.
84. M. SCHELLEKENS, o.c., 623.
85. Directive 96/9/EC of 11 March 1996 on the legal protection of databases, OJL 077, 27 March 1996, art. 11.
86. Apis-Hristovich EOOD v Lakorda AD, C-547/07, 5 March 2009, paragraph 55.
87. British Horseracing, paragraph 70.
88. British Horseracing, paragraph 71.
89. High Court of Justice, 9 February 2001, the British Horseracing Board/William Hill Organization (HC 2000 1335).
90. Tribunal de Grande Instance de Paris, 5 September 2001, Keljob/Cadremploi (00/17196).
91. British Horseracing, paragraph 72.
92. Article 7(5) Database Directive.
93. British Horseracing, paragraph 89.
94. Article 9(b) Database Directive. Not all Member States (e.g., Italy and France) have implemented this exception.
95. For a general overview of copyright issues in scientific research (and an argumentation why it has become increasingly difficult for researchers to make free use of research material), see J. REICHMAN and R. OKEDIJI, "When Copyright Law and Science Collide: Empowering Digitally Integrated Research Methods on a Global Scale," Minnesota Law Review 96, 4, 2012, http://papers.ssrn.com/sol3/papers.cfm?abstract-id=2149218.
96. See the report of G. WESTKAMP, The implementation of Directive 2001/29/EC in the Member States, Part II, February 2007, goo.gl/ Iea1EJ, 11.
97. F. JENNINGS and J. YATES, "Scrapping over data: are the data scrapers' days numbered?," Journal of Intellectual Property Law & Practice 4, 2, 2009, 125.
98. Contra: Cour d'appel de Pau (France), 23 March 2012, goo.gl/a5s04e, in which the Court of Appeal emphasized that a clause needs to be presented in a very explicit manner, while in casu Facebook's terms and conditions were very long, and written in English.
99. With respect to the templates used for diagnostic tests to determine cognitive impairment, see E. ROSATI and C.M. ROSATI, "Copyright in diagnostic tests: not yet the end for fair use," Journal of Intellectual Property Law & Practice 7, 7, 2007.
100. A. STROWEL (Belgium) distinguishes between artistic works, functional works and factual works. For the latter category, the creative freedom is limited, because the contents often substantially constraints the expression form. See A. STROWEL, "L'originalité en droit d'auteur: un critère à géomé trie variable", J.T. 1991, as from 513.
101. This distinction was inspired by an American report: see Office of Technology Assessment, Intellectual Property Rights in an Age of Electronics and Information, Washington, 1986.
102. Infopaq I, paragraph 47.
103. M. VAN EECHOUD, "Along the Road to Uniformity e Diverse Readings of the Court of Justice Judgments on Copyright Work," JIPITEC 2012, 1, paragraph 63.
104. For example, according to the Spanish Supreme Court, Google's public cache service cannot rely on the exception to offer copies of web pages. See Spanish Supreme Court April 3, 2012, Sentencia n 172/2012, http://pdfs.wke.es/8/ 6/1/5/pd0000078615.pdf.
105. M. BORGHI and S. KARAPAPA, Copyright and Mass Digitization, Oxford, Oxford University Press, 2013, 135-138.
106. Editions du Seuil et autres v Google Inc et France, Paris District Court, 18 December 2009.
107. Opinion of Advocate General Stix-Hackl, C-338/02, 8 June 2004, paragraph 38.
108. See Fixtures Marketing, paragraph 27: "the resources used for the purpose of giving the database its function of processing information, that is to say those used for the systematic or methodical arrangement of the materials contained in that database and the organisation of their individual accessibility".
109. See the discussion in section 3.4 below.
110. Compare with the fourth factor in the fair use doctrine in the United States, where courts do not need to consider markets that are only potential. See Sony, 446 U.S. at 456, as cited by the court in the HathiTrust case at 14.
111. See www.creativecommons.org.
112. F. JENNINGS and J. YATES, "Scrapping over data: are the data scrapers' days numbered?" Journal of Intellectual Property Law & Practice 4, 2, 2009.
113. For the Netherlands, see Hoge Raad (Supreme Court), 22 March 2002, NVM v De Telegraaf (scraping in competing real estate websites e not allowed);
114. Rechtbank Alkmaar, 7 August 2007, Jaap (scraping in competing real estate websites e allowed if texts and thumbnails were sufficiently small).
115. See pending case C-202/12, regarding the Dutch "Gaspedaal" cases (car sales website), where a competing website sent search queries from users to the original website, and merely temporarily showed the search result. One of the questions submitted to the CJEU is whether 100,000 queries per day conflicts with the normal exploitation or legitimate interests. Also in the Netherlands, see Gerechtshof Arnhem, Zoekallehuizen.nl, 4th July 2006, available at goo.gl/RO2BLt, in which the argument was raised that the additional load due to the indexing process was negligible e an argument that was not disputed by the other party.
116. A. BEUNEN and T. SCHIPHOF, Legal aspects of web archiving from a Dutch perspective (Report commissioned by the National Library in The Hague), October 4, 2006, http://repub.eur.nl/res/ pub/11343/KB-Legal-Aspects-WebArchiving-EN.pdf
117. 17 Contra: see the input from various publishers in HOUSE OF COMMONS, The Hargreaves Review of Intellectual Property: Where next? e First Report of Session 2012-13, available at goo.gl/GqeUkJ, 15-18.
118. See the Acceptable Use Policy at about.pinterest.com/use/.
119. See article 3 of the "Statement of Rights and Responsibilities" www.facebook.com/legal/terms, the robots.txt file (facebook.com/ robots.txt), as well as the Automated Data Collection Terms. Facebook does allow some of the popular search engines to crawl its site.
120. See section 3 of Tumblr's Terms of Service at www.tumblr.com/policy/en/ terms-of-service.
121. See goo.gl/zitW.
122. See Landgericht Hamburg, case 308 0 248/07, sec. B.I.6.d. Even so, as discussed below, the German Supreme Court eventually allowed thumbnails in search engines on the basis of an "implied license".
123. I. HARGREAVES, Supporting Document T to the Hargreaves report e Text Mining and Data Analytics in Call for Evidence Responses, May 2011, www.ipo.gov.uk/ipreview-doc-t.pdf, 6.
124. According to the outline Presentation (goo.gl/igzgXv) of the Working Group 4 on text and data mining (installed by the European Commission to discuss this topic), the majority of researchers and research institutions pointed out that mining of texts for which a general license was already obtained, should not be subject to an additional license or payment. In their perspective, the "right to read includes the right to mine".
125. R. VAN DER NOLL, S. VAN GOMPEL, L. GUIBAULT, J. WEDA, J. POORT, I. AKKER and K. BREEMEN, "Flexible Copyright," August 2012, 20.
126. Copiepresse v Google Inc, Court of Appeal Brussels, 5 May 2011, English translation available on goo.gl/mIhtlF.
127. See the French case in Editions du Seuil et autres v Google Inc et France, Paris District Court, 18 December 2009.
128. For a general discussion of the Google Books project, see W.B. CHIK, "The Google conundrum: Perpetrator or facilitator on the net? e Forging a fair copyright framework of rights, liability and responsibility in response to search engine 2.0 e Part II: The Google Books Search Project," Computer Law & Security Review 27, 4, August 2011.
129. Vorschaubilder I, BGH, 29 April 2010, Az. 1 ZR 69/08;
130. Vorschaubilder II, BGH, 19 October 2011, Az. 1 ZR 140/10.
131. In Copiepresse v Google, the first court (13 February 2007, Auteurs & Media 2007, 107) even stated explicitly (chapter 8) that it was not the copies in the internal cache that were being contested, but their accessibility for the public.
132. R. VAN DER NOLL et al., "Flexible Copyright," August 2012, 23.
133. For example: the original elements of the image remain available in the smaller image; the author's right to integrity may be breached due to the reduction in size; even the mere act of creating the smaller image entails a temporary copy for which the applicability of the exception must be assessed. See Y. VAN DEN BRANDE, "Scrapen: het verzamelen van on-line publiekelijk beschikbare gegevens," Auteurs & Media 2009, 2009, 609.
134. Vorschaubilder I, BGH, 29 April 2010, Az. 1 ZR 69/08;
135. Vorschaubilder II, BGH, 19 October 2011, Az. 1 ZR 140/10.
136. The implied license theory also upholds in the US copyright system (as part of the fair use doctrine), and is even argued to align with the historical roots of the copyright system: see O. FISCHMAN AFORI, "Implied License-An Emerging New Standard in Copyright Law," Santa Clara Computer and High Technology Law Journal 25, 2008;
137. J.S. SIEMAN, "Using the implied license to inject common sense into digital copyright," NCL Rev. 85, 2006.
138. Spanish Supreme Court April 3, 2012, Sentencia n 172/2012, http://pdfs.wke.es/8/6/1/5/pd0000078615.pdf.
139. See M. PEGUERA, "Copyright Issues Regarding Google Images and Google Cache", in A. LOPEZ-TARRUELLA (ed.), Google and the Law e Empirical Approaches to Legal Aspects of Knowledge-Economy Business Models, The Hague, T.M.C. Asser Press, 2012, 193.
140. For a general discussion, see R. XALABARDER, "Spanish Supreme Court rules in favour of Google search engine", Kluwer Copyright Blog, available at goo.gl/lHJJfc.
141. For a general discussion under Belgian law, see J. VELAERS, Bronnen en beginselen van het recht, University of Antwerp, 2013, 156-161;
142. for French law, see P. MALINVAUD, Introduction à l'étude du droit (12e édition), LexisNexis, 2009, 149-155.
143. More in particular the question of whether a coutume contra legem can take priority over Community legislation (exhaustive list of exceptions in the Copyright Directive). This view is contested, although supported by some authors: see J. CARBONNIER, Variations sociologiques: mélanges Levasseur, 1992
144. e G. CORNU, Introduction, 13e é d., nr. 263.
145. I. HARGREAVES, Digital opportunity: A Review of Intellectual Property and Growth, May 2011, page 47, available at www.ipo.gov.uk/ipreview-finalreport. pdf.
146. See the December 2012 publication on "Modernising Copyright: A modern, robust and flexible framework" (available on www.ipo.gov.uk/ response-2011-copyright-final.pdf).
147. The current state of the implementation of the Hargreaves report can be consulted at www.ipo.gov.uk/types/hargreaves.htm
148. and, specifically with respect to the new data mining exception for non-commercial research, www.ipo.gov.uk/techreview-dataanalysis.pdf.
149. "Non-consumptive Use," Discussion Paper on Copyright and the Digital Economy, Australian Law Reform Commission, www.alrc.gov.au/publications/ copyright-and-digital-economy-dp-79.
150. 17 USC x107.
151. In the same vein: R. VAN DER NOLL et al., "Flexible Copyright," August 2012, 7.
152. P.N. LEVAL, "Towards a Fair Use Standard", Harvard Law Review 103 (1990) 1105.
153. P.B. HUGENHOLTZ and M.R.F. SENFTLEBEN, Fair use in Europe. In search of flexibilities, November 2011, 6;
154. M. SENFTLEBEN, "Bridging the Differences between Copyright's Legal TraditionseThe Emerging EC Fair Use Doctrine," Journal of the Copyright Society of the USA 57, 3, 2010, 527.
155. K. PATEL, "Authors v. Internet Archives: The Copyright Infringement Battle over WEB Pages" J. Pat. & Trademark Off. Soc'y 89, 2007, 416.
156. A.S. DIAZ, "Fair use & mass digitization: the future of copydependent technologies after Authors Guild v. Hathitrust," Berkeley Tech L.J. 2013, July.
157. In Kelly v. Arriba Soft Corp (280 F.3d 934, 9th Cir. 2002), the defendant stored copies of images (including photos of the claimant) into its internal databases, and showed these images to end-users as thumbnails. The court applied the fair use doctrine, because it found a transformative use of the photos, as the thumbnails served a different function than the original photos. Even though the search engine showed each photo in its entirety this was found to be reasonable, considering the function of the thumbnails. The thumbnails were also not deemed to cause any market harm to the photographer.
158. Google's thumbnails were also discussed in the Perfect 10 v Amazon and Google case (508 F 3d 1146, 9th Cir. 2007). There, the claimant argued it had a market for thumbnails, because it licensed another company to use smaller sizes of its images on mobile phones. Even so, the appeal court was not convinced that the thumbnails interfered with this market, and instead focused on the transformative use of the photos.
159. Field v. Google (2006 WL 242465, D. Nev. 2006), in which Google was sued for providing cached links to copyrighted texts.
160. A.V. v. iParadigms, 544 F. Supp. 2d 473, 2008.
161. Authors Guild Inc v HathiTrust, No 11 Civ 6351 (HB), 2012 US Dist. For a general discussion, see A.S. DIAZ, o.c. It concerned a case where a consortium of university libraries scanned and stored entire copies of works in order to provide full-text search capabilities. According to the court, HathiTrust's index was a transformative use that did not serve the same expressive purposes as the underlying copyrighted works.
162. I. HARGREAVES, Supporting Document T to the Hargreaves report e Text Mining and Data Analytics in Call for Evidence Responses, May 2011, www.ipo.gov.uk/ipreview-doc-t.pdf, 1.
163. For a legal-philosophical overview of this issue, as well as a description of the legal treatment of machine-reading in copyright history, see M. BORGHI and S. KARAPAPA, Copyright and Mass Digitization, Oxford, Oxford University Press, 2013, 45-56.
164. Hargreaves report, Supporting Document EE, Economic Impact of Recommendations, available on www.ipo.gov.uk/ ipreview-doc-ee.pdf, 34.
165. This uncertainty should, however, not be exaggerated: the closed catalogue of exceptions in the EU approach also gives rise to numerous disputes. See P.B. HUGENHOLTZ and M.R.F. SENFTLEBEN, Fair use in Europe. In search of flexibilities, November 2011, 7-9.
166. P. VAN EECKE, "Online service providers and liability: a plea for a balanced approach," Common Market law review, 2011.
167. M. SENFTLEBEN, "Bridging the Differences between Copyright's Legal TraditionseThe Emerging EC Fair Use Doctrine," Journal of the Copyright Society of the USA 57, 3, 2010, 540.
168. This aligns with the conclusion of a working group on text and data mining, which argues that "at present, European researchers and technology companies are mining the web at legal and financial risk, unlike their competitors based in the US, Japan, Israel, Taiwan and South Korea who enjoy a legal limitation and exception for such activities." e Open letter, available at goo.gl/yUX56.
169. P.B. HUGENHOLTZ and M.R.F. SENFTLEBEN, Fair use in Europe. In search of flexibilities., November 2011;
170. R. VAN DER NOLL et al., "Flexible Copyright," August 2012, 7.