One of the less-noticed Hargreaves Review's proposals, supported by the Government in its Response to the Hargreaves Review, is to widen the copyright exception for non-commercial research to cover both text and data mining.
What is text mining?
A good explanation appears in the National Centre for Text Mining's Submission to Hargreaves. "The overall goal of text mining is to help discover new knowledge from old, based on processing unstructured data, i.e. the text of scientific articles, books, reports etc. It applies a battery of natural language processing techniques to identify and extract information from such content, turing this information into structured data, over which data mining algorithms are run to find interesting, previously unknown associations."
Does European Copyright law allow the UK to introduce an exception for data mining?
Yes. The Copyright Directive allows member states to keep/ introduce an exception for scientific research on tw conditions: (1) the author has to be attributed unless this turns out to be impossible and 2) to the extent justfied by the non-commercial purpose. That is mirrored in the existing 'fair dealing' exception for research in the UK's Copyright, Designs & Patents Act 1988, as amended.
So what's the issue?
Data mining requires access to publishers' databases and therefore depends on licences to use material protected by copyright and database right. Those in favour of a broader exception argue that publishers' licences that permit data mining are too narrow or the terms unclear. Publishers reject this.
So what's the solution?
Text mining is a vital tool in research by revealing new patters and trends in data. But an over-broad exception would undermine publishers' legitimate licensing and business models.
The Government needs some joined up thinking here. It's advocating the creation of a Digital Copyright Exchange to turbocharge rights licensing and trading. Surely we can develop automated licensing solutions which accomodate non-commercial research covered by exceptions for free, whilst charging for commercial exploitation?
Licensing, not overbroad exceptions, is the right way forward.
Have a good week,