The Stationers' Company

Effect since implementation

​

The implementation of the exception for non-commercial TDM meant that there was no longer a need for non-commercial researchers (including those engaged in universities and other third- level institutions) to request permission to perform TDM on any works to which they have lawful access. This then increases the potential for TDM, as it is no longer as arduous or time- consuming as it would have previously. The elimination of the requirement to search for each individual copyright-holder and request permissions made TDM a more viable research method for a variety of researchers. This means that, as awareness of TDM grows, more researchers will consider it a viable research method, leading to the discovery of new information and greater analysis of the vast swathes of data available in the digital age. Although publishers no longer have any control over whether or not their content may be mined for data analysis purposes, they maintain the option to deny any access to the work at all – researchers are required to have ‘lawful access’ to the work before they may copy it for TDM purposes.73 Publishers are also allowed to impose restrictions in order to maintain the stability or security of their network (eg by limiting downloads to a certain number per month) but these measures must not unreasonably restrict researchers’ access to the works.74 The exception also had the effect of denying publishers the chance to monetise non-commercial TDM requests. However, it is not outside the bounds of imagination that publishers may simply incorporate the costs of supporting non-commercial TDM into their licence fees for access, given that the right of access includes the right to mine.

With regard to commercial exploitation of TDM techniques, this remains the province of individual publishers. Thus, they may create or decline to create systems which will allow commercial researchers to mine content, provide APIs, partake in joint initiatives such as

​

71 ibid reg 3(1)(a).
72 See, for example, The Copyright and Rights in Performances (Quotation and Parody) Regulations 2014, SI 2014/2356, reg 3.
73 CDPA (n 35) s 29A(1).
74 IPO, ‘Exceptions to Copyright: Research’ (2014) <https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/375954/Research.pdf > accessed 25 March 2015, 6.

195

CrossRef75 or PLSclear TDM,76 or simply refuse all TDM applications entirely. However, given the proportion of TDM requests which were non-commercial and now no longer need to be supported, it is not outside the bounds of possibility that the implementation of the non- commercial TDM exception has disincentivised publishers and content owners to create systems which can deal with requests for commercial TDM permissions in a timely and efficient manner. Nonetheless, as TDM continues to grow, it is in publishers’ best interests to develop procedures that can efficiently issue TDM permissions – this may take the form of publisher- specific initiatives, or joint projects between multiple publishers. It is important, however, that publishers and rights holders are allowed to retain the freedom to control the commercial exploitation of their works – including permitting or denying copying for the purposes of data mining and analysis. To do otherwise would be to undermine the integrity of copyright rules entirely.

One area of potential confusion which may arise is due to the fact that there is no clear delineation of the distinction between commercial and non-commercial research given in the legislation. While a common-sense approach would seem to be the most likely approach which will be taken, it is an area which may yet lead to a certain amount of litigation before the courts as to what exactly constitutes non-commercial research, and to what degree such researchers can rely on the TDM exception. For example, a university researcher who conducts research using TDM techniques is likely to fall within the exception, especially if the results of their research are published in a scientific journal, usually for no fee. In fact, if the research is published under certain Open Access protocols, the researcher may even be required to pay a fee in order to publish their research.77 This is clearly not a commercial gain for that researcher. However, if their research then forms a chapter of a monograph, textbook, or other larger work for which royalties are payable, does this change the nature of the research? While the research may be the same, the method of publication in the latter case accrues a certain amount of financial gain. Thus, is the research no longer eligible to take advantage of the exception? This is a question which will have to be answered through litigation, as it is not explained in the legislation itself. One possibility is that a researcher who did not need to seek permissions to mine content originally would then have to retroactively seek a licence if they wished to

​

75 CrossRef (n 58).
76 PLSclear TDM (n 60).
77 For more on Open Access, see Chapter 7.

196

subsequently publish for commercial reasons, but this would be particular to individual circumstances.

The exception is relatively new, and has not yet been subject to litigation. In the event of commercial research attempting to rely on the TDM exception, it would be interesting to consider which aggrieved party would take a suit against the commercial researcher. For each individual author, if the amount of data analysed was large, then the degree of use of their copyright would be so minute that it may not accrue more than nominal damages. In that case, then, the most likely parties to take a case against commercial miners would be publishers or licensing societies. Much like the NLA, the licensing society for newspapers in the UK, was the body which took a case against news clippings services,78 it would not be outside the bounds of imagination to suggest that licensing societies would again be the most logical avenue for litigation against TDM infringement. However, this of course, is conjecture, and remains to be seen.

Naturally, TDM requests are not limited solely to publishers – there are data resources which could provide great insights and scientific developments if mining of the content were permitted. For example, while mining the scientific literature and published articles about the use of a specific drug could provide new data on potential side effects, a much wider source of information on this could be found in the NHS patient records. Therein one would find a comprehensive record of the effects of a particular drug on numerous patients and, were mining of that content allowed, new uses for drugs could be discovered with relative ease. However, the mining of data sources other than those which are commercially available (ie those to which a subscription may be purchased) is something outside of the ambit of the current exception. Thus, it does nothing to enable the progression of knowledge in areas other than specifically non-commercial TDM with data which may be lawfully purchased or obtained, and therefore fails to account for much data which could be the source of new scientific insights. This use of TDM may well be one which is impossible to license, and thus could be a potential case for a second TDM exception, but it is one which must be carefully researched and defined before legislation is enacted.

​

78 NLA Media Access, ‘Newspaper Websites – Copyright Law’ (2014) <http://www.nlamediaaccess.com/uploads/public/News/Summary%20of%20Legal%20Proceedings%20 -%20Background%20for%20Journalists%20_June_2014.pdf>accessed 12 January 2016.

197

In August 2015, the UK Publishers Licensing Society conducted a survey of its members regarding the use of the TDM exception.79 Of those publishers who responded to the survey, less than 20% stated that they had received mining requests.80 The survey covered the number of requests to mine which were received in the year 2014, and the overall number of requests from 111 publishers was just 91.81 Although there is no data for why this number is so low, the PLS suggested that it may have been due to licence terms including permissions to mine content, and the use of automated systems, such as CrossRef, which allow researchers to access material for TDM purposes without having to contact the publishers directly.82 This low instance of requests for TDM permissions was not expected to rise greatly in the future, nor had it seen a notable rise or fall with the implementation of the UK TDM exception. It is suggested that the existence of automated services such as CrossRef, PLSclear TDM, and the Copyright Clearance Center, may well have negated the need for a TDM exception at all.83

However, it is important to note that as of the date of conclusion of the research project, an independent research project had not been completed on the effect of the TDM exception in the UK. It is wise to note that figures reported by an organisation such as the PLS may be open to accusations of bias, and thus should be treated with some scepticism, and an independent project would be advisable for greater reliability.

​

Calls for wider exception in Europe

​

The movement toward TDM was not isolated to the UK. Across the EU, there was also a realisation of the growing need to acknowledge TDM as a research method, and consider the need for legislation to regulate the use of TDM. This resulted in the creation of a TDM working group in the Licences for Europe initiative, which was established in December of 2012 and launched the following February.84 Unfortunately, the working group did not proceed as successfully as had been hoped. In February 2013, a number of participants of the working group – mostly librarians – walked out, on the basis that the scope of the group was not wide enough to adequately account for the possibilities of TDM. They set out their concerns in a letter to the

79 Publishers Licensing Society, ‘Survey shows text and data mining supported by licensing not copyright exceptions’ (PLS News and Events, August 2015) <http://www.pls.org.uk/news-events/n-tdm- august-15/> accessed 18 October 2015.

80 ibid.
81 ibid.
82 ibid.
83 ibid.
84 Michel Barnier, ‘Licences for Europe: quality content and new opportunities for all Europeans in the digital era’ (Speech at the Launch of the initiative ‘Licences for Europe’, 4 February 2013) <http://europa.eu/rapid/press-release_SPEECH-13-97_en.htm> accessed 18 December 2015.

198

relevant Commissioners (Barnier, Geoghegan-Quinn, Kroes, and Vassilou) later that same month, arguing that the group considered only re-licensing content, and not a wider variety of solutions to the TDM quandary, including exceptions or limitations to copyright law.85

Nonetheless, the plenary group produced a pledge, one of the ten pledges for getting more content online which were published in December of 2013.86 This stated that the group had achieved ‘Easier text and data mining of subscription-based material for non-commercial researchers: a commitment by scientific publishers’ in the form of a proposed licensing clause which allowed for TDM at no extra cost for non-commercial research. At the date of publication of the ten pledges, thirteen publishers had signed up to include this licence term.87

Despite the pledges, this did not appear to be a sufficient movement towards enabling TDM, and thus in 2014 a veritable plethora of reports were commissioned and published by two European Directorates-General (DGs), focusing on TDM. DG Research and Innovation published the Hargreaves-led Expert Group’s report ‘Standardisation in the area of innovation and technological development, notably in the field of Text and Data Mining’,88 while DG Market commissioned and supported March’s De Wolf and Partners ‘Study on the legal framework of text and data mining (TDM)’.89 A third report from 2014, ‘Assessing the economic impacts of adapting certain limitations and exceptions to copyright and related rights in the EU:

​

85 – –, ‘Letter from participants in response to “Licences for Europe- A Stakeholder Dialogue” text and data mining for scientific research purposes workshop’ (February 2013). Signatories were Sara Kelly, Executive Director, The Coalition for a Digital Economy Jonathan Gray, Director of Policy and Ideas, The Open Knowledge Foundation John McNaught, National Centre for Text Mining, University of Manchester Aleks Tarkowski, Communia Klaus-Peter Böttger, President, European Bureau of Library Information and Documentation Associations (EBLIDA) Paul Ayris, President, The Association of European Research Libraries (LIBER) Brian Hole, CEO, Ubiquity Press Ltd. David Hammerstein, Trans- Atlantic Consumer Dialogue, and the letter was additionally supported by more than fifty other interested parties.

86 Licences for Europe, ‘Ten pledges to bring more content online’ (2013) <http://ec.europa.eu/internal_market/copyright/docs/licences-for-europe/131113_ten-pledges_en.pdf> accessed 18 December 2015, 8.
87 ibid 8.
88 The Expert Group, ‘Standardisation in the area of innovation and technological development, notably in the field of Text and Data Mining’ (2014) <http://ec.europa.eu/research/innovation-union/pdf/TDM- report_from_the_expert_group-042014.pdf> accessed 18 December 2015.
89 Jean-Paul Traille, Jérôme de Meeûs d’Argenteuil and Amélie de Francquen, ‘Study on the legal framework of text and data mining (TDM)’ (2014) <http://ec.europa.eu/internal_market/copyright/docs/studies/1403_study2_en.pdf> accessed 12 January 2016 (De Wolf and Partners).

199

Analysis of specific policy options’ from Charles River Associates (CRA), was also supported by DG Market.90

While these three reports were all published in the same year, they came to markedly different conclusions about the need for a TDM exception, and indeed the wider copyright system. While the Expert Group report from DG Research and Innovation called for an exception to copyright law, this was only a precursor to its recommendation to entirely overhaul the European copyright system. The De Wolf study from DG Market focused on the current legal framework, and whether TDM would fall under any of the exceptions which are currently mandated under European law. It concluded that TDM could not be permitted under any of the existing exceptions, and recommended that a TDM exception be put into place to remove ‘unjustified obstacles’91 to data analysis. The report by CRA, also DG Market, offered four ways to proceed regarding the question of TDM and relevant legislation, before recommending the implementation of a non-commercial exception for scientific research in circumstances where licences were not offered. This exception could be relied upon where it was not possible to obtain a TDM licence, but would cease to apply once a publisher began to offer TDM licences, thus incentivising publishers to create licensing terms in order to avoid falling subject to the exception.

The Expert Group92 which was responsible for the DG Research report was led by Professor Ian Hargreaves, who was also responsible for the 2011 Hargreaves Review,93 discussed earlier in the chapter. Within the Expert Group report, one finds the statement that prolific use of TDM would add tens of billions of Euros in value to the EU’s aggregate GDP, but the important thing to note here is that this estimate is not based on empirical research. Indeed, ‘[t]here are no empirical estimates of the impact of TDM on the productivity effect of research’,94 and the

​

90 Julian Boulanger and others, ‘Assessing the economic impacts of adapting certain limitations and exceptions to copyright and related rights in the EU: Analysis of specific policy options’ (2014) <http://ec.europa.eu/internal_market/copyright/docs/studies/140623-limitations-economic-impacts- study_en.pdf> accessed 12 January 2016 (Charles River Associates).

91 DeWolf and Partners (n 85) 116.
92 The Expert Group was assembled for the purposes of the report discussed in this chapter. It was composed of Professor Hargreaves, of Cardiff University, Dr Lucie Guibault (University of Amsterdam, the Netherlands), Dr Christian Handke (University of Amsterdam and Erasmus University, the Netherlands), Professor Peggy Valcke (KU Leuven, Belgium) and economist Bertin Martens (JRC, IPTS, Seville). They were supported by Dr Ros Lynch (Department for Business Innovation & Skills, United Kingdom) as rapporteur. It further received research assistance from Dr Sergey Filippov, Assistant Professor of Innovation Management at Delft University of Technology and Non-resident Fellow of the Lisbon Council.
93 Hargreaves (n 12).
94 Expert Group (n 84) 33.

200

figures proffered in the report were based on data from a JISC report95 which applied only to the UK, and did not take into account either the variation between different European countries’ research outputs or the unique position that the UK occupies as a research centre. Nonetheless, the Expert Group report then went on to use the JISC figure (of 2%) and apply this to the EU as a whole, to increase the real value of EU research output by €5.3 billion. It used an elasticity estimate from Guellec and Van Pottelsberghe’s 2004 OECD economic study96 to model the long- term impact of a change in the volume of R&D expenditure, which came out at a €32.5 billion increase in terms of innovative products, productivity and consumer welfare increases.97 It did not, however, state over what length of time this long-term increase would apply. In addition, this would be a minor increase in terms of proportion of EU GDP – only 0.26%.

On the back of these assertions, the report proposed three action points, the first of which was licensing initiatives.98 The second was a wide-ranging exception for TDM, which removed it entirely from the scope of European copyright and database law.99 This was, however, an interim suggestion, to be implemented pending the progress of the paper’s third suggestion, a large- scale reform of European copyright law.100

The Expert Group dismissed almost out of hand the possibility of licensing options, citing the needs of digital age researchers, who ‘require legally reliable research access to many types of database, spread across numerous media platforms, disciplines, organisations and countries’.101 While this is not untrue, to dismiss the possibility of licensing is to deny rights holders a level of control over their protected works, and also fails to take into account joint enterprises such as the aforementioned PLSclear TDM engine and CrossRef.

With regard to a TDM exception, the report came to some of the same conclusions that will be put forward later in this chapter. The distinction between commercial and non-commercial research can, at times, be hard to make. Further, the definition of ‘scientific’ is also difficult to ascertain. The report concluded that the only viable exception would be one which applied to all scientific researchers, both commercial and non-commercial, in order to avoid confusion,

95 ibid 25.
96 Dominique Guellec and Bruno Van Pottelsberghe, ‘From R&D to Productivity Growth: Do the Institutional Settings and the Source of Funds of R&D Matter?’ (2004) CEB Working Paper No 04/010. 97 Expert Group (n 84) 33.
98 ibid 65.
99 ibid 66.
100 ibid 67.
101 ibid 65.

201

Chapter 6: The Hargreaves Exceptions – A Case Study

and to improve the status quo.102 However, the distinction between copying expressive works and the incidental copying which occurs as part of the data analysis process would still be difficult to draw. It is on the basis of this difficult distinction that the report’s made its ultimate recommendation – a new Copyright Directive.103 Thus, while it recommended the implementation of a ‘specific and mandatory exception to remove text and data mining for scientific purposes from the reach of European copyright and database law’,104 this was only to be considered as a medium-term amelioration pending the new copyright Directive.

This recommendation, however, was incomplete – it was a recommendation that did not take into account the nuances involved in scientific research. While the distinction between scientific non-commercial research and commercial research can be difficult to make, to create an exception which applied to all scientific research would be heavy-handed and would not take into account the distinction in purpose behind commercial and non-commercial research. Further, the recommendation of the exception also failed to consider the crucial forerunner of whether or not the researching party would have lawful access to the material to be analysed. To allow TDM without lawful access to material would be overreaching on the part of the legislature, which, rather than enhancing the value of European scientific research, could undermine it entirely. As has been established many times throughout this thesis, copyright walks a balance between access and incentive, and distinctions such as lawfully owning a content of work to be analysed would be a vitally important one in maintaining that balance.

The second study,105 by De Wolf and Partners,106 assessed the legal framework of TDM in Europe. It considered the possibility that data analysis could fall under the current copyright and database legislation, before concluding that this would largely be impossible.107 It did point out that TDM, in part, could be justified under existing exceptions, as a part of non-commercial scientific research, but the lack of a specified exception could lead to confusion for researchers. It stated very clearly that more research would be needed, due to the fact that TDM as a research tool was so new.108 It then moved on to consider the possible construction of a new European

​

102 ibid 68.
103 ibid 68.
104 ibid.
105 De Wolf and Partners (n 85).
106 De Wolf and Partners is a Belgian/Luxembourgish legal firm that specialises in corporate, commercial, TMT, tax, employment and real-estate law. It often produces high-level European reports, including in the area of intellectual property reform.

107 De Wolf and Partners (n 85) 41-84. 108 ibid 96.

202

data mining exception, taking guidance from existing legislation in the InfoSoc Directive,109 and the Database Directive110 and their provisions on both copyright and sui generis rights.111 The recommendation of the report was that, if a TDM exception were to be implemented, it should apply to non-commercial research which is mainly scientific in purpose, and that any requirement to acknowledge the source should be left to individual data miners.112 The report further suggested, inter alia, that such an exception should be made mandatory across all member states, unlike the InfoSoc exceptions, as this would lead to greater harmonisation in the EU, and would also simplify the possibilities of cross-border research,113 and lastly suggested that it should not be possible to override the exception by contract, much like the UK exceptions discussed earlier in this chapter.114

The De Wolf and Partners report was cautious in its approach, and extensive in its analysis, both of the current law, and the potential for change. Its suggestions were built on the need for further research into the area, and considered also the need for balance in the implementation of any new exceptions. This author disagrees with the need for a mandatory pan-European exception, especially given that the scientific research exception in InfoSoc has not been uniformly implemented, but respects that this would lead to greater clarity and more ease of cross-border research across the EU.

The third report, by Charles River Associates (CRA),115 was not limited solely to considering TDM. Rather, it considered a variety of possible exceptions to copyright law, including remote access, private copying, TDM, and library e-lending. It considered the licensing context in the EU at the time, and pointed out that for each possible TDM project, the time input in contacting the rights holder in order to seek permission to mine would be considerable.116 As well as this, many publishers had not yet developed TDM licences and thus the costs associated with obtaining a TDM licence could vary. This uncertainty could then make undertaking mining projects difficult, time-consuming and intimidating. The study made a strong point, however,

​

109 InfoSoc Directive (n 1).
110 Directive 96/9/EC of the European Parliament and of the Council of 11 March 1996 on the legal protection of databases [1996] OJ L77/20.
111 De Wolf and Partners (n 85) 99.
112 ibid 100.
113 ibid 107.
114 ibid 106-113.
115 Charles River Associates is a global consulting firm, headquartered in Boston, which offers economic, financial and strategic consulting to major law firms, corporations, accounting firms, and governments around the world.
116 Charles River Associates (n 86) 16-17.

203

in pointing out that the market for TDM licences was nascent – publishers could not be expected to have developed licences to feed a demand that did not exist only a few short years previously.117 It further pointed out that there was no empirical evidence which stated the demand for TDM licences.118 However, the market was already adapting, as initiatives which have already been mentioned, such as PLSclear TDM, CrossRef’s automated TDM initiative, and the commitment by scientific, technical and medical (STM) publishers to a standard TDM licence indicated.119

The study assessed four separate policy options for TDM:

1. Maintaining the status quo for TDM

2. A specific harmonised and mandatory exception for text and data mining for thepurpose of non-commercial scientific research in the absence of a licence agreementenabling text and data mining

3. A specific harmonised and mandatory exception for text and data mining for thepurpose of non-commercial scientific research

4. A specific harmonised and mandatory exception for text and data mining for thepurpose of scientific research (both commercial and non-commercial)120

Each of these four options was individually assessed, with the research concluding that options 3 and 4 (both of which were non-overridable exceptions) would not be warranted.121 It is worth noting that option 3 would be an extension to the entire EU of the situation which was in place in the UK since the implementation of the new TDM exception in 2014. The report made the point that, especially with regard to an exception which was both commercial and non- commercial, removing the potential for financial gain from TDM would then disincentivise publishers from developing the TDM-specific structures which would allow TDM to flourish, and further benefit the EU as a whole.122 This exception, which at first glance seemed nothing but positive, could then lead to more difficulty in utilising TDM techniques. With regard to option 1, the report pointed out that maintaining the status quo could leave high transaction costs in place. It did acknowledge that new developments were emerging and maintaining the status quo could still encourage investment in TDM infrastructure.123 It did not state that

​

117 ibid 70.
118 ibid 65.
119 ibid 70.
120 ibid 74-81. 121 ibid 78, 80. 122 ibid 72.

123 ibid 74.

204

maintaining the status quo would be impossible, or even undesirable, but maintained that it would not be the most preferable outcome for Europe.124

The recommendation which received highest priority was the implementation of a specific harmonised and mandatory exception for TDM for the purpose of non-commercial scientific research in the absence of a licensing agreement enabling TDM.125 This exception, option 2, in contrast to options 3 and 4, could be over-ridden by the implementation or offering of a licence by a rights holder. This would apply only where there was lawful access to the content being mined – ie where the researcher or research institute had a subscription to a particular database. This option would still allow publishers to provide licences for TDM, but would not prevent researchers from accessing content where a publisher did not offer licences for TDM. It was suggested in the study that this would have little to no effect on larger publishers, as they had established licences for TDM (particularly for scientific purposes) and thus the exception would not apply.126 It was further suggested that the exception would also encourage smaller publishers to amend their subscription agreements in order to add mining clauses, meaning that the effect on them would be limited only for as long as they had not yet included or provided such a licence. It pointed out also that content from smaller publishers was less frequently mined than that from larger publishers.127 The study lastly suggested that the loss of revenue by publishers affected by the exception would encourage them to develop licences, leading to lower transaction costs in the future.128

The detriment felt by the implementation of option 2 over maintaining the status quo would be limited, in that it would likely affect only smaller publishers, and even then only until they produced their own TDM licences. The increased production of TDM licences, the report suggested, would then likely reduce transaction costs.129 The recommendation would ensure that researchers were able to mine content, even where there was no licence available, but also maintain rights holders’ incentives, allowing the market the freedom to set its own rates. This would be an improvement over the status quo, as it would allow rights holders to develop their own TDM clauses, but also require that permission to mine be made available to researchers.

124 ibid.
125 ibid 81. 126 ibid 75. 127 ibid 75. 128 ibid 75-6. 129 ibid 76.

205

Chapter 6: The Hargreaves Exceptions – A Case Study

This minimal level of intervention would allow the market to develop its own solutions in time, while still providing encouragement for rights holders.130

Of course, the discussion of TDM has not been limited solely to reports commissioned by legislators. Other interested parties have also been making their opinions known, across Europe. By way of example, in July 2014 the French Council for the Protection of Literary and Artistic Property (CSPLA)131 published the report of their ‘Mission sur l’exploration de données’,132 proffering their opinion that legislative initiative on TDM would be premature, and the market should be encouraged to develop its own licensing solutions.133

The Mission report, written by Jean Martin and Liliane de Carvalho, gave first a description of TDM, with specific reference to the subject matter concerning the CSPLA (that is to say, artistic and literary works), pointing out that the area had been slower to adapt than scientific journals, and had adapted differently – the notion of Open Access meant that for 50% of journal articles, mining was already permitted.134 The report then went on to note that the authorisation of rights holders, without legislative intervention, was still required for mining activities, presenting a logistical problem where the quantity of objects to be mined was great.135 However, it also noted that the value of TDM was more than that of a parasite – TDM could potentially create and add further value, justifying the use of TDM techniques.136 Thus, it suggested, a balance must be struck between the often intimidatingly high transaction and administrative costs and the value that TDM could create in the future. It also praised the Licences for Europe movement, mentioned earlier in this chapter, which created a standard TDM clause, along with commitments to implement this licence by multiple publishers.137

The discussion of TDM in this report is as extensive as those commissioned by the European DGs, but considered from a different angle. The final product of the report was that it made twelve recommendations, aimed at various different aspects of TDM, from changing the

130 ibid 77.
131 More correctly, ‘Conseil supérieur de la propriété littéraire et artistique’, but translated to English for ease.
132 Jean Martin and Lilian de Carvalho, ‘Mission sur l’exploration de données’ (July 2014) <http://www.culturecommunication.gouv.fr/Politiques-ministerielles/Propriete-litteraire-et- artistique/Conseil-superieur-de-la-propriete-litteraire-et-artistique/Travaux-du- CSPLA/Missions/Mission-du-CSPLA-relative-au-text-and-data-mining-exploration-de-donnees> accessed 30 March 2015.

133 ibid 4-5.
134 ibid 10.
135 ibid 16.
136 ibid 18.
137 ibid 6.

206

perception of TDM from parasitic to symbiotic, to creating a public policy designed to manage the specific rights required for mass use of works for TDM, including prioritising self-regulation over legislative intervention, allowing access to public data for TDM, and maintaining that non- legislative conviction within the European Union and WIPO structures.138 The report also recommended that a two-year deadline be set to prepare an industry report on the possibility of an eventual need for legislative intervention, but maintained that legislative intervention at the point of publication (July 2014) would be premature. A move towards legislation is easy to implement, but hard to retract, and the recommendations of the CSPLA report, which included a review in two years, would allow the market enough time to adapt to what was still a very new development, while showing a willingness to co-operate with miners, reusers, and content adaptors in the future. The CSPLA report was indicative of the willingness of rights holders to engage with content miners, and come up with solutions to allow greater use of TDM, but also demonstrates the reluctance to embrace a mandatory legislative exception to TDM – this would remove all incentives for rights holders to cooperate in any way, and may then damage the prospects of TDM in the future.

Another viewpoint worth considering is that of librarians. As mentioned earlier in this chapter, librarian representatives walked out of the 2013 Licences for Europe working group on TDM, as they felt their concerns were not being heard.139 The European Association of Research Libraries, (LIBER )140 continued to campaign for a European TDM exception, including in their response to the 2014 European Consultation on Copyright.141 In February of 2015, former LIBER president Paul Ayris met Commissioner Oettinger, the Commissioner for Digital Economy and Society, to campaign for a mandatory, pan-European, non-overridable TDM exception to European copyright law, citing issues of cross-border work and damaging the European single market if this were to be left to individual Member States to implement, like the InfoSoc copyright exceptions.142

The variety of opinions available from interested parties, librarians, researchers, content owners and independent research bodies meant that the European Parliament had a fine line to walk

​

138 ibid 4-5.
139 Letter from participants (n 81).
140 LIBER actually stands for Ligue des Bibliothéques Européenes de Recherche.
141 LIBER, ‘LIBER Response to the Public Consultation on the review of the EU copyright rules’ (2014) <http://libereurope.eu/wp-content/uploads/Public%20Copyright%20Consultation%20Response.pdf> accessed 30 March 2015.
142 Paul Ayris, ‘LIBER Argues For Pan-European TDM Exception’ (LIBER, 23 February 2015) <http://libereurope.eu/blog/2015/02/23/liber-argues-for-pan-european-tdm-exception/> accessed 30 March 2015.

207

in order to reach a solution acceptable or at least tolerable to all interested parties, which would continue to promote European growth, research and profitability. The benefits of effective TDM are difficult to quantify in either financial or public good terms, but there seemed to be little disagreement that there was distinct potential. Thus, on such a hot-button topic, Europe was under an intense amount of pressure to find a solution which would result in maximised benefits through the effective use of TDM, and it seems that there is no easy answer to this quandary.

​

Conclusion

​

Text and data mining is a research tool that has the potential to speed up research in all areas of life, with possible untold benefits for all interested parties. However, enabling free use of TDM is fraught with issues, as it must walk a line between the interests of researchers and content owners, while preserving the free market. There are many issues that could arise with the implementation of a TDM exception, which may need some guidance before the exception could come into effect.

For example, there may be difficulty in defining what exactly scientific research is, and it is not suggested where such a definition or distinction should come from. The layman’s understanding of scientific research would suggest testing of a hypothesis within the hard sciences – so-called for their quantifiable data produced, testable and reproducible via the scientific method, generally understood as being the natural sciences (astronomy, biology, chemistry, earth science, and physics). Of course, this definition could also be expected to include STEM subjects (science, technology, engineering, and maths) or STM (scientific, technical, and medical) research. It is clear that even within the constraints of hard science, there is the possibility for variation between subject areas. Nonetheless, the layman’s understanding fails to take into account the possibility to open to expand the term scientific research to its broadest possible interpretation, which could encompass also soft sciences, which could include social science, political science, and psychology.143 Social science in particular is an expansive area, encompassing economics, history, anthropology, law, linguistics, geography, education, and more subject areas.144 A narrow definition of scientific research would prevent many university and third-level non-commercial researchers from taking advantage of any potential exceptions or initiatives for scientific research, which would, in all likelihood, be a misinterpretation of the

​

143 For a discussion of some of the differences between hard and soft sciences, as well as a mention of the sometimes artificial distinctions between them, see Larry V Hedges ‘How Hard is Hard Science, How Soft is Soft Science?’ (1987) 42(2) American Psychologist 443.
144 Norman W Storer, ‘The Hard Sciences and the Soft: Some Sociological Observations’ (1967) 55(1) Bull Med Libr Assoc 75.

208

intention behind any exception. Furthermore, as the De Wolf and Partners report pointed out, limiting a TDM exception to solely scientific research would prevent many projects which have as their main aim scientific research from completing the entirety of their project, as secondary objectives which may veer more toward market research would not fall under the exception.145

A second problem to be tackled would be the issue of delineating between commercial and non- commercial research. Such a distinction would likely be difficult to draw, and guidance would need to either be vague, in the form of general principles, or very specific – even still, it is certainly not outside the bounds of imagination that a lack of clarity would lead to researchers relying on an exception which may not apply to them, if they are ‘fringe’ cases.

The suggestion of implementation of an exception for TDM could be considered as premature. According to guidance from the Green Book, exceptions to copyright law should only be implemented where there is a market failure.146 While the market for TDM is nascent, it has not failed entirely. It has not yet been given time to grow. There are multiple initiatives in place to provide licences for mining, both individually by publishers,147 and through larger co-operative initiatives.148 There is thus no immediately obvious need for an exception for publishers, and indeed implementing an exception could lead to further issues in distinguishing between scientific and non-scientific, and commercial and non-commercial research. Furthermore, rights holders are price discriminating, as stated in the CRA study – they often grant TDM licences free of charge to non-commercial researchers149 and as the UK PLS survey stated, only a small proportion of publishers were receiving TDM requests, due to services such as PLSclear TDM and CrossRef and the inclusion of TDM clauses in standard licence agreements.150 This then alleviated the difficulty of licensing which was faced by potential TDM users in previous years.

The suggestion made by multiple parties, including De Wolf and Partners and LIBER, that a potential exception should be mandatory for all EU Member States would lead to a scenario in which all member states would have an exception for TDM, but not for scientific research in general, as there is not yet harmonisation of all the InfoSoc exceptions across the EU.151 This

​

145 De Wolf and Partners (n 85) 62.
146 Green Book (n 3) 11.
147 For example Elsevier has a mining API, and multiple publishers have signed up to the Licences for Europe standardised mining clause.
148 PLSclear TDM (n 60) and CrossRef (n 58), for example.
149 Charles River Associates (n 86) 64.
150 PLS (n 75).
151 Eleonora Rosati, ‘Copyright in the EU: in search of (in)flexibilities’ (2014) 9(7) JIPLP 585.

209

scenario would then mean that certain countries would be put in the situation where TDM is permissible, but other aspects of scientific research are not, a strange scenario by any reach of the imagination.

Finally, the increase in Open Access (OA) means that a large proportion of scientific research is already available under a licence which allows reuse and transformative uses.152 In those circumstances, a TDM exception is redundant, as it does not place any onus on the rights holders to allow anything which they had not already allowed. Furthermore the UK requires that journal articles and conference proceedings published after 1 April 2016 must be made OA in order to qualify for the Research Excellence Framework, and encourages all institutions to implement an OA policy in advance of this date.153 This means that most, if not all, university and third-level researchers will make their research available OA online, negating the need for a TDM exception in order to mine this content. Similar policies may well be implemented across other parts of the EU, as the move towards OA is gaining momentum worldwide.154

As the non-commercial TDM exception in the UK has already been implemented, the issues of defining scientific research and non-commercial purposes may well be laid out for consideration in future case law. The existence of the UK TDM exception can act as a model, example, or even cautionary tale for the implementation of a wider European TDM exception. We can see from the implementation of this UK exception that there may well be no need for a European-level initiative to implement a new TDM exception, as it is possible to frame a TDM exception within the exceptions already permitted under the 2001 InfoSoc Directive (in the UK’s case, under the exception for Research and Private Study).155 Thus, the immediate implementation of a new instrument creating a TDM exception would be redundant. Any move toward a greater TDM exception in the EU should be approached thoughtfully and through a dialogue between stakeholders on all sides of the debate. The possibility of implementation of TDM exceptions in Member States under current legislation means that there is no need to rush to bring in a pan- European TDM exception, and thus legislators can afford the luxury of time, discussion, and careful consideration before deciding whether or not such an exception is a necessity.

​

152 Martin and de Carvalho (n 128) 10 suggested that more than 50% of journal articles were open access. 153 HEFCE, ‘Policy for open access in the post-2014 Research Excellence Framework’ (July 2014) <http://www.hefce.ac.uk/media/hefce/content/pubs/2014/201407/HEFCE2014_07.pdf> accessed 30 March 2015.

154 For more, see chapter 7.
155 Of course, it is possible that this exception may be judicially reviewed, like the Private Copying exception, but this remains to be seen.

210

In conclusion, while there may be criticisms made of the UK TDM exception, it may in the future prove to be a good test case in order to allow for greater consideration and careful planning of any potential implementation of a greater TDM exception in the EU as a whole. The wide variety of opinions and expert reports available on the topic of TDM, as well as the strongly-held and research-supported views of interested parties, including rights holders, researchers and other related parties such as libraries, mean that any legislative intervention on the topic of TDM in the EU should be carefully considered, using the UK as an example of how things may pan out in an EU environment. There is little doubt that TDM is an exciting and potentially game- changing research tool which may redefine the landscape of scientific research, but to heedlessly update copyright law without careful consideration of the delicate and nuanced balance which is required between rights holders and researchers could have adverse consequences which would counteract the financial and welfare benefits that TDM could offer. Furthermore, as a variety of research techniques, and applying to a range of types of data for assessment, TDM runs into further issues with regard to privacy, data protection, ownership of data, and myriad other issues which cannot be resolved with a blanket exception.

For these reasons, the author feels that a pan-European exception for TDM would be premature, and could detrimentally affect the development of TDM in Europe. Thus, empirical research should be conducted and carefully assessed before changes are made to the European copyright structure. This should focus on the possible implications of a mandatory exception against an optional exception, whether the exception applies to non-commercial only or all research, scientific, mainly scientific or all research, and whether it can be excluded through contract. The current framework allows individual member states to implement TDM exceptions if they so wish, and those who choose to do so will have the example of the UK to use as a framework on which to base their own exceptions. The author believes that the introduction of non- commercial scientific research exceptions for TDM across EU member states could be beneficial for all parties involved, both on the rights holder and researcher side, but without empirical evidence, this is only an opinion.

Furthermore, while TDM exceptions as suggested and discussed in this chapter may well improve access to scientific literature which is available through database and journal publication, there is a wealth of privately held data which would likely be more difficult to access, and thus could be a better case for a TDM exception. The lawful access requirement for the current TDM exception in the UK means that only data which is available for purchase can be mined. Thus, a carefully framed exception which improves access to other datasets, such as NHS patient records, may also be a candidate for consideration. Of course, given that this would

211

also interact with issues of privacy and data control, it would be a more involved and lengthy process than an exception which contemplates only data which is available for purchase. Ultimately, though, it would be an exception which would be more beneficial not only to researchers, but potentially also to humanity as a whole.

While the UK TDM exception can be used as a test case, on the other hand, the implementation of the private copying exception can be used as a demonstration of how not to implement an exception. Despite the example of several other European countries’ private copying levy schemes, the UK declined to include a similar levy scheme. They also employed flawed research and impact assessments, which wrongly denoted the amount of harm to rights holders which would result from such an exception. As such, despite the thirteen-year gap between the ability to implement the exception156 and its actual beginnings,157 and the almost-four years from the publication of the Hargreaves Review158 to the private copying exception’s debut, this can still be seen as a cautionary tale as to the wisdom of failing to properly research the implications of certain exceptions. As a result, while there is still a market failure regarding private copying, there is now no private copying scheme in place in the UK, and no intention to rectify this.159 Thus, in the future, the UK government must be sure to obtain objective and verifiable evidence of the effect of implementing exceptions to copyright before proceeding, lest the exception be quashed less than a year after its coming into effect.

156 Via the powers of the InfoSoc Directive (n 1), which was 2001.
157 The exception came into effect on 1 October 2014 (n 14).
158 The Hargreaves Review (n 12) was published in May of 2011.
159 Pinsent Mason, ‘UK government scraps plans to legalise private copying’ (Out-Law.com, 18 November 2015) <http://www.out-law.com/en/articles/2015/november/uk-government-scraps-plans-to-legalise- private-copying/> accessed 18 December 2015.

​