Text/Data Mining Potential Legal Issues Naomi Korn IP Consultant naomi@naomikorn.com Disclaimer: The contents of this presentation are for general advice and best practice purposes only. They do not constitute legal advice and the author can give no assurances, warranties regarding the accuracy, currency of applicability in relation to specific circumstances. In such circumstances, professional legal advice should be sought. This presentation is licensed under a Creative Commons Attribution Share Alike Licence 3.0 http://creativecommons.org/licenses/by-sa/3.0/ How High Can You Jump? • Current legislative environment makes it nearly impossible to legitimately carry out the processes necessary for text/data mining without permission or ideally an exception • Strong likelihood of legal issues occurring in association with all the processes associated with data/text mining • At almost every intersection along the text/data journey, there are legislative and/or licensing potholes and pitfalls that disrupt opportunities CONTENT SUPPLY: Contractual Restrictions • Contractual clauses associated with the supply of articles etc can place restrictions on the user • Authors themselves may still assign rights or grant exclusive licences which preclude them and/or their organisations from text/data mining without additional permissions • No provisions to prevent contractual clauses from over-riding copyright exceptions even if we were to have an exception to text/data mining • Not abiding by contractual terms = Potential breach of contract and infringement of copyright • Even open access journals can present issues in terms of the particular CC licence selected. I.e. licences with a “ND” restriction can prevent adaptation and mining; CC licences in general (apart from CC Zero) can lead to Attribution stacking if many articles mined. FORMAT OF CONTENT • Format may preclude data/text mining i.e. PDF AND/OR • technical protection measures may be used to lock up access to content (reinforced by contractual terms) • NB: Remove of any Technical Protection Measures = Infringement of copyright COPYING OF CONTENT: Legislative Restrictions CDPA Section 28 A Temporary copies exception is solely concerned with incidental and immediate copying But: Any copying which is a “consumption” of the work, temporary or not, requires permissions from the rights holder “…the exception is concerned with transient copies which have no value in themselves and which do not prejudice the rights holder by interfering with the normal exploitation of the work…” NLA v Meltwater (Court of Appeal) [2011] EWCA Civ 890 http://www.bailli.org/cgibin/markup_cgi?doc=/ew/cases/EWCA/Civ/2011/890.ht ml THIS MEANS THAT EVEN IN THE PROCESS OF A MACHINE MAKING A COPY OF AN ARTICLE ETC FOR THE PURPOSES OF TEXT MINING, AN INFRINGEMENT IS TAKING PLACE BECAUSE THERE IS INHERENT VALUE IN THE ULTIMATE COPYING OF THE WORK! PRESENTATION OF DATA • Copyright infringement if more than just facts? • Presentation of collation of facts = Infringement of Database Right? • Even though unlikely infringement of copyright in data mining, maybe issues associated with breach of contract as well as potential database rights arising What do we need? • Exception for text/data mining BUT ALSO: – We need a holistic approach to legislative change as recommended by Professor Hargreaves in this report – Exception for data/text mining will be useless without: • Provisions to prevent contract law from over-riding copyright exception • Provisions to prevent TPMs from over-riding copyright exceptions • Provisions to facilitate copying for preservation purposes • Solutions for Orphan Works