UK Government consultation on AI & copyright: text and data mining


On 29 October 2021, the UK Intellectual Property Office issued an open consultation to examine certain aspects of copyright and patent law in the context of AI. The consultation follows in light of views expressed in response to its Call for Views on AI and IP.

One particular aspect of the consultation concerns text and data mining and, in particular, the potential approaches to permit text and data mining activities without requiring authorisation of a copyright owner. Presently, the UK has a very narrow copyright exception for text and data mining that enables computational analysis solely for the purpose of non-commercial research.

Five potential approaches are outlined in the consultation, namely:

  1. Maintain the current text and data mining exception and make no change.
  2. Maintain the current text and data mining exception but seek to improve the licensing environment to facilitate better access to data.
  3. Expand the current exception to enable commercial research and mining of databases.
  4. Expand the current exception to include text and data mining generally but allow copyright owners to ‘opt out’ works which they do not wish to be mined. This is broadly equivalent to the exception introduced by the EU copyright directive (which the UK did not implement following Brexit).
  5. This approach would be the same as 4 but with no opt out.

Mining is one way of obtaining data to train AI models, but presently, in the UK, the current exception for text and data mining has such a narrow scope that anyone wishing to do so is likely to face a very real and significant risk of infringing copyright. While there are some ways to mitigate this legal risk, if the UK is to achieve its stated desire to be among the “global AI superpowers” it is hard to see how it can maintain the current exception.

In some respects, this issue is a microcosm of a much larger balancing exercise being played out across Europe between the interests of IP owners and data holders on the one hand and those of data users on the other. Ostensibly, there is now a policy priority favouring the accessibility and sharing of data (reflected in both the UK’s National Data Strategy and the EU Commission’s European strategy for data) and as such the UK may ultimately be tempted to adopt a more permissive regime.

The consultation closes at 23:45pm on 7 January 2022.