Skip to content

Text and data mining copyright exception proposal to encourage AI focus and innovation in the UK



The UK IPO is proposing to include a new exception in UK copyright law to permit data mining and analysis of datasets for commercial use.

From 29 October 2021 to 7 January 2022, the UK Intellectual Property Office (UK IPO) ran a consultation on AI and intellectual property. One of the key topics covered in the consultation was text and data mining using copyright material. Text and data mining (TDM) involves the use of computational techniques to analyse large amounts of data to establish trends and examine various analytics. TDM is used to train AI systems and is also used extensively in various fields such as research, journalism, marketing, and by cultural heritage organisations.

The results of the consultation (Response to Consultation Report) were published on 28 June 2022. The Response to Consultation Report notes that the UK IPO plans to introduce a new copyright and database exception which allows TDM for any purpose. This proposed change will have an impact on anyone licensing or permitting third parties to access their data. The intention behind the change is to encourage the development of wider innovation and AI-driven businesses, given their reliance on large datasets to generate machine learning.

The new exception, if implemented, would go considerably beyond the current text and data analysis exception (under section 29A(1)(a) Copyright, Designs and Patents Act 1988), in two key respects:

  • The exception would relate not only to copyright, but would also entail an explicit exception to the sui generis database right which exists separately under UK law, and which protects databases from the unauthorised extraction or reutilisation of the whole or a substantial part of the contents of a database; and

  • It would permit TDM for commercial purposes, whereas the existing exception requires that the purpose of TDM must be for non-commercial research.

The new exception would preserve the requirement for anyone copying a work (or extracting or reutilising the contents of a database) for the purposes of TDM to have “lawful access” to the work (or database). The Response to Consultation Report suggests that this will mean that rights-holders will continue to have safeguards in place to protect their content. Lawful access could include, for example, subscription-based access or permissions agreed between parties through terms and conditions.

The new exception will, like the current TDM exception, provide that any contract term purporting to override the exception (and prevent or restrict the acts permitted by the exception) would be unenforceable. In other words, it will not be possible to contract out of the exception.

In the UK IPO’s response to calls for views (Calls for Views Report) which sets out the response to its call for views, it noted the various concerns held by rights-holders. Some commentators pointed to the fact that that any new exception allowing more permissive use would shift the balance unfairly against creators. The Calls for Views Report noted that many rights-holders accept the rationale for the current exception of TDM for non-commercial purposes, but there are clear concerns that an exception for commercial purposes could unfairly prejudice their legitimate interests and tip the balance against rights-holders.

Many aspects of the UK IPO’s proposal remain unclear and this article examines some of the key grey areas.

International obligations

Like all UK copyright law exceptions, the TDM exception will have to comply with the UK’s remaining international obligations, including those under international copyright treaties, notably Article 9 of the Berne Convention. Article (2) states that the reproduction of literary and artistic works can be reproduced in certain special cases provided that this reproduction does “not conflict with a normal exploitation of the work and […] not unreasonably prejudice the legitimate interests of the author”.

Article (2) is normally understood as a safeguard to ensure that exceptions to copyright do not undermine normal economic exploitation, such as by means of licensing. For this reason exceptions under copyright law are generally restricted to use for non-commercial purposes, unless there is a clear rationale for permitting a commercial use in limited circumstances (for example, reporting current events).

Some commentators argue that the right of reproduction granted by the new exception is limited and would not necessary fall foul of the Berne Convention. It has been argued by commentators that TDM is not a reproduction but would constitute a separate category of copying which does not fall within the Article 9 definition of the Berne Convention. It is worth remembering that the TDM exception is not designed to permit all copying, but is confined to copying for the purpose of conducting computational analysis of anything recorded in the work. It might also be argued that the lawful access requirement preserves the economic interests of right-holders.

The UK IPO states that data protection is outside of the scope of the consultation. The TDM exception will need to be considered separately under data protection law by entities who will be required to understand the relationship between the UK GDPR and TDM exception. The UK Government has expressed its future plans to replace the UK GDPR “with our own business and consumer-friendly, British data protection system”. The intention to better service AI and technology businesses therefore appears to underpin the proposed reforms to both copyright and data protection law.

Other rights-holder restrictions

As already mentioned, not all rights-holders consider this proposed exception to benefit creators, despite the consultation’s emphasis on the safeguards in place to protect rights-holder content. These safeguards include the requirement for lawful access, which allow the rights-holder the option to charge for access to their data, restrict access to it or even to withhold access altogether. For example, rights-holders could elect to implement subscription barriers so that access to their works would only be granted in exchange for payment.

Organisations wishing to apply TDM to a database would need to have a licence from the owner of the copyright in the database. It seems reasonably clear that such a licence could not prohibit TDM, as such a term would be unenforceable (see above). The Response to Consultation Report has not otherwise addressed what restrictions rights-holders could impose without being unenforceable, but it would seem to be the case that the only option for a rights-holder, if it wishes to prevent TDM of its content, is either to refuse to grant access to it, or ensure that its scale of charges reflect use of that content for TDM.

What is not clear is whether a rights-holder could permit access under licence for a standard fee, but require a further payment for TDM. Arguably this would be an (albeit indirect) contractual restriction on TDM and potentially unenforceable.

Sui generis right for databases

There are two types of intellectual property protection for databases: sui generis database right (known simply as “database right”) and copyright. Both rights are automatic, unregistered rights that allow the owner to control certain uses of their databases.

Database right protects the contents of a database. To qualify, there needs to have been a substantial investment in obtaining, verifying, or presenting the data. The acts restricted by database right are the unauthorised extraction or reutilisation of the whole or a substantial part of the contents of a database.

As noted above, the existing TDM exception relates only to copyright and not to sui generis database rights. Having said that, there is an exception to database right which applies to “fair dealing” by a lawful user of the database, where the contents have been extracted for the purpose of illustration for teaching or research and not for any commercial purpose. To some extent, this aligns with the existing exception for TDM.

The proposed TDM exception would explicitly permit the extraction and reutilisation of the contents of a database for the purposes of TDM by a lawful user of a database, whether those purposes were commercial or non-commercial. This is bound to impact the value of already established sui generis database rights, but the UK IPO has not shared any details of the interplay between the two intellectual property regimes at this stage.

General comments

Overall, it is striking how the UK IPO is clearly favouring one industry at the expense of another. As currently proposed, the balance is set to tip drastically in favour of licence-holders and AI-developers. The UK IPO appears to have side-lined the concerns of rights-holders in order to prioritise being at the forefront of AI-driven development and technology. At this stage, many rights-holder concerns remain unanswered.

Read our update article “Status update on the proposed new copyright law exception to permit text and data mining for commercial use” here.

This publication is a general summary of the law. It should not replace legal advice tailored to your specific circumstances.

© Farrer & Co LLP, November 2022

Want to know more?

Contact us

About the authors


Sufi Rahimi


Sufi advises across a broad range of sectors, having joined from a media, technology and retail background. She enjoys working on a blend of IP advisory, data protection and commercial contract work and adds value for clients through her ability to advise across these disciplines in parallel.

Sufi advises across a broad range of sectors, having joined from a media, technology and retail background. She enjoys working on a blend of IP advisory, data protection and commercial contract work and adds value for clients through her ability to advise across these disciplines in parallel.

Email Sufi +44 (0)20 3375 7491
Back to top