Legal framework of textual data processing for Machine Translation and Language Technology research and development activities/Concluding Remarks

Concluding RemarksEdit

What has been clearly shown in this report is that LR processing for purposes of MT & MP is a complex issue that requires a number of permissions to be obtained in order to occur at the lowest possible transaction cost. At this stage the following options are in place:

  • Unrestricted use on the basis of copyright and data protection exceptions: this is the best possible solution, as it incurs the lowest possible transaction costs. However, the operation of such a regime is far from being a reality. The legal uncertainties and inconsistencies are such that render most of such exceptions practically inapplicable.
  • Use of standard licensing regimes allowing open access, such as the one provided through the Creative Commons and various Privacy Commons projects. This is a solution that provides legal certainty, but is dependent upon voluntary action and is, hence, fragmented and not uniform in its application.
  • Requesting ad hoc licences: this is the legally safest solution but has the greatest transaction cost and is not certain that it will lead to the identification of the rights-holders or the obtaining of the necessary permissions.

Overall this report proposes to take a clear and unambiguous position towards the introduction of positive user rights

  • at a global level, i.e. through WIPO,
  • at a regional level, i.e. through and EU Directive and
  • at a national level by supporting positive user rights policies, where these exist (e.g. Public Sector Information rights).

Using open or extended collective licences has been a key element of a more balanced copyright policy for the past ten years. However, it is high time to move forward seeking change at the legislative rather than the contractual level. Open licensing has been tremendously successful in terms of making the business case for open data and content, but much less successful in terms of reducing clearance costs and uncertainty inherent in a copyright system that requires no formalities or registration for protection and that still relies in the EU context on an exclusive set of limitations and exceptions rather than a fair use doctrine.

What ten years of experience with open content licences has amply demonstrated is that users are no more passive recipients of content but rather active creators, especially in the context of language resources, and, hence, they require positive rights rather than mere exceptions, particularly when such exceptions are list- and not doctrine- based.

The re-use of LRs is not something that may be easily solved through licensing, since it poses an excessive clearance and licence choice cost upon the Language Technologies providers, it does not solve the problem of orphan works and does not take into consideration the issue of works belonging to rights-holder other than the LR providers.

Hence, the overall direction needs to shift and have three successive goals:

  1. the overarching goal should be the establishment of positive user rights
  2. if this is not possible, the aim should be the transition to a fair use doctrine in the EU
  3. if this is not immediately achievable, extensive collective licensing and linking of funding with open licensing should be sought

While licensing remains an option, it is the very last option we should consider. It currently serves Member States with developed documentation of their works and a positive IPR trading balance, but is highly unlikely to offer any substantial benefits to Member States with nascent information society services and negative IPR trade balances.