Reed Smith In-depth

Key takeaways

  • UK government is consulting on the introduction of a new "text and data mining" exception to copyright modelled on the EU law exception adopted in 2019.
  • The exception would apply for any purpose, including commercial use, but only where the user has lawful access to the works and  would be subject to the significant caveat that right holders could opt out through an agreed mechanism.
  • As in the EU AI Act, the exception would also be subject to transparency mechanisms about training material sources.
  • The consultation also discusses AI outputs, noting that the existing copyright framework is suitable for addressing infringing outputs. It suggests removing the "computer-generated works" provision from the CDPA, supports AI output labelling obligations, and considers the contentious issue of inference and  personality rights.
  • Responses to the consultation can be submitted by 25 February 2025 with evidence encouraged and submissions being made public.

In brief

  • Two years after abandoning its proposal to create an unconditional text and data mining (TDM) exception to copyright, the UK Intellectual Property Office (IPO) is consulting again on copyright and TDM.
  • In a carefully worded text released just before Christmas, prefaced by the government stressing the importance of both the creative industries and the AI sector to the UK economy, the consultation skirts around some of the most controversial issues by focusing on catching up with the EU approach to TDM and suggesting practical ways of addressing some of its side effects.
  • The consultation provides limited discussion of the important controversies underpinning the TDM and copyright debate, opting instead to focus on the purported benefits of the proposed approach, which would include: (i) enhancing right holders’ control of their material and their ability to be remunerated for its use; (ii) supporting wide access to high-quality material to drive the development of leading AI models in the UK; and (iii) securing greater transparency from AI developers to build trust with creators, creative industries and consumers.
  • The government has introduced a proposal for a new TDM exception having the following key features: (i) it would apply to TDM for any purpose, including commercial use; (ii) it would apply only where the TDM user has lawful access to the works; (iii) right holders could opt out of the exception “through an agreed mechanism” (machine-readable when content is available online); and (iv) it would be underpinned by high-level transparency about the sources of training material.
  • The text recognises the challenges faced by right holders in knowing who is entitled to rights and how to validly reserve them. It considers several methods, including the robot.txt standard, metadata or online registries, and also invites respondents to share their views on the role of collective licensing.
  • The second half of the consultation shifts its focus to AI outputs and its key takeaways can be summarised as follows:
    • The existing copyright framework is well-suited to address copyright infringing outputs.
    • The rules regarding the copyrightability of AI-generated content follow EU and U.S. law save that AI “entrepreneurial works” – such as synthetic sound recordings, films, broadcasts and published editions – are entitled to protection.
    • The “computer-generated works” provision under the Copyright, Designs and Patents Act 1988 (CDPA) could be removed.
    • The government’s text supports the introduction of AI output labelling obligations.
    • The introduction of personality rights is not the primary focus of the consultation but is on the table as an option.
    • The contentious topic of private copying is returning to the agenda.
  • Views on the proposals and issues contained in the consultation can be submitted via Citizen Space or by email by 25 February 2025. The government welcomes submission of evidence, including case studies and data. Respondents should be aware that summary responses will be made public and should advise when submitting their responses whether any information therein is confidential or sensitive. However, full responses may be published on request pursuant to access to information regimes, following individual assessment.