Consultation outcome

Artificial Intelligence and Intellectual Property: copyright and patents: Government response to consultation

Updated 28 June 2022

Executive summary

1. Artificial intelligence (AI) is a transformative technology, which is already revolutionising many areas of our lives. The Taskforce on Innovation, Growth and Regulatory Reform (TIGRR) report urged UK leadership on AI. Unleashing the power of AI is a top priority in the plan to be the most pro-tech Government ever. The National AI Strategy sets out how the Government aims to secure the UK’s position amongst the global AI superpowers.

2. AI can support innovation and creativity in a range of ways. It can be a tool for scientists, entrepreneurs, and artists, enabling new human inventions and creations. Some believe that AI will soon be inventing and creating things in ways that make it impossible to identify the human intellectual input in the final invention or work. Some feel this is happening now.

3. Intellectual property (IP) gives researchers, inventors, creators, and businesses the confidence to invest their time, energy and money in doing something new. UK business invests more than £130 billion a year in knowledge assets. IP rights protect around £63 billion of this. These assets are vital to the industries that bring us the innovation and products that add value to our lives.

4. The Government sought evidence and views on a range of options on how AI should be dealt with in the patent and copyright systems. This response sets out the conclusions of that process.

5. We considered three specific areas:

• copyright protection for computer-generated works (CGWs) without a human author

• licensing or exceptions to copyright for text and data mining (TDM), which is often significant in AI use and development

• patent protection for AI-devised inventions

6. For computer-generated works, we plan no changes to the law. There is no evidence at present that protection for CGWs is harmful, and the use of AI is still in its early stages. As such, a proper evaluation of the options is not possible, and any changes could have unintended consequences. We will keep the law under review and could amend, replace or remove protection in future if the evidence supports it.

7. For text and data mining, we plan to introduce a new copyright and database exception which allows TDM for any purpose. Rights holders will still have safeguards to protect their content, including a requirement for lawful access.

8. For AI-devised inventions we plan no change to UK patent law now. Most respondents felt that AI is not yet advanced enough to invent without human intervention. But we will keep this area of law under review to ensure that the UK patent system supports AI innovation and the use of AI in the UK. We will seek to advance AI inventorship discussions internationally to support UK economic interests.

Introduction

9. Unleashing innovation and creativity will be at the heart of the post-pandemic recovery and support British businesses to build back better. Innovation and creativity are also central to the Government’s levelling up plans.

10. The National AI Strategy sets out how the Government aims to secure the UK’s position amongst the global AI superpowers. This includes how the Government will invest in the long-term needs of the UK’s AI ecosystem. Every sector and region can then benefit from the transition to an AI economy by focusing on encouraging innovation whilst protecting the public.

11. The UK Innovation Strategy: leading the future by creating it sets out our long-term plan for delivering innovation-led growth. Its primary aim is to boost private sector investment across the whole of the UK. This will ensure the right conditions for all businesses to innovate and give them the confidence to do so.

12. Intellectual property (IP) gives researchers, inventors, creators, and businesses the confidence to invest their time, energy and money in doing something new. It underpins economic growth by incentivising investment, safeguarding assets and enabling the sharing of know-how in technologies like artificial intelligence (AI). The benefits of IP are not only for businesses’ bottom lines. There is investment in talented workers, regional clusters, the people, and communities who benefit from new ideas, products and entertainment. It also helps the nation’s standing on the world stage.

13. AI has huge potential to rewrite the rules of whole industries. It can drive economic growth and transform all areas of life. The UK is a global superpower in AI, well placed to lead over the next decade - a genuine research and innovation powerhouse. Our work on AI and IP supports the ambitions set out in the National AI Strategy. The strategy will ensure the UK continues as a world leading location for research, development, commercialisation, and deployment of AI.

14. We looked at the interaction between IP and AI in Artificial Intelligence and intellectual property: call for views. Several outcomes were identified as a result, including three areas that should be the subject of further consultation. These were covered in Government’s consultation of 29 October 2021 on how copyright and patent law can best support AI development and deployment. We now set out the Government response to that consultation.

About the consultation

15. This consultation sought evidence and views on a range of options on how the patent and copyright systems should deal with AI technology.

16. We consulted on three specific areas:

• copyright protection for computer-generated works without a human author. These are currently protected in the UK for 50 years. But should they be protected at all and if so, how?

• licensing or exceptions to copyright for text and data mining, which is often significant in AI use and development

• patent protection for AI-devised inventions. Should we protect them, and if so, how should they be protected?

17. There are two overall objectives. The first is to encourage innovation in AI technology and promote its use for the public good. The second is to preserve the central role of intellectual property in promoting human creativity and innovation. To support these objectives, the consultation aimed to collect the best available economic evidence.

18. We held six roundtables in November and December 2021. Participants included individuals, large and small businesses, and a range of organisations. Contributors were from sectors including the creative industries, tech industry, pharmaceuticals, the third sector, academia, and legal and IP professions. More than 60 people participated in the roundtables and shared their experiences and thoughts. There were also 88 written submissions received. We also talked to AI service/equipment providers, artists, entrepreneurs, inventors, academics and other IP offices.

Copyright in computer-generated works

Summary

19. Computer-generated works (CGWs) are copyright works without a human author. They are currently protected in UK copyright law. As part of our consultation, we asked whether they should continue to be protected and, if so, how. We also asked about related issues: the impacts of similar provisions in designs law, and risks of false attribution.

20. 61 written responses expressed a view on one or more of these issues. The majority favoured no change to the law (Option 0). Fewer respondents supported either of the other two options. Many, on all sides, said that there is little evidence of protection for CGWs having significant impacts at present.

21. We have decided to adopt Option 0: make no changes to the law. There is no evidence at present that protection for CGWs is harmful, and the use of AI is still in its early stages. As such, a proper evaluation of the options is not possible, and any changes could have unintended consequences. But we will keep the law under review and could amend, replace or remove protection in future if the evidence supports it.

Options and responses

21. We received a broad range of written responses, including from trade associations in the publishing, technology, and music industries, academics, libraries, and cross-sector bodies.

22. Most respondents agreed that protection for CGWs does not seem to be used widely at present. Several noted that there has been only one court case dealing with CGWs and this did not involve AI. Some respondents said that users and developers of AI do rely on these protections. There was little supporting evidence for this.

Option 0: make no changes

24. Many of those supporting Option 0 said that it would be premature to make changes. They said that there is no evidence that existing protection causes harms. But the longer-term implications of removing it are unclear. They also said doing so would remove legal certainty, while leaving the law unchanged offers stability. Some respondents who favoured Option 0 conceded that, as AI and its uses develop, this issue might need to be looked at again. Alternative approaches could become suitable in future.

Option 1: remove protection for CGWs

25. Those who supported Option 1 included academics, research centres, and libraries. These respondents argued that the protection was not necessary or beneficial. Some raised concerns that ‘originality’, a necessary condition for protection, is unclear in the case of CGWs. Others noted that AI systems can generate huge numbers of works. This means that protecting CGWs could hamper innovation and crowd-out human creators and their works. They argued that removing protection for CGWs would promote and protect the status of human creativity.

Option 2: replace protection for CGWs with a new right of reduced scope or duration

26. Some respondents supported Option 2. This was also the second preference of many who supported Options 0 and 1. They argued that a new right could be crafted to incentivise investment in AI, whilst still promoting human creativity. Most of them suggested the right should have significantly shorter duration than copyright (such as two to ten years). But many respondents said that more evidence would be needed to decide on the scope and duration of a new right.

Other CGW related issues

27. We also asked about the implications of AI and our policy on CGWs on protection for designs. Few responses addressed this question. No significant concerns were raised about the implications of protecting AI-generated material on the designs system.

28. Many respondents gave views on the risks of CGWs being falsely attributed to human authors. This might be for the extra term of protection available to human-authored works, or for reputational reasons. Several respondents raised concerns about this issue. But most considered that existing law would be adequate. In particular, respondents highlighted laws on passing off, fraud, privacy, and the need to establish authorship in cases of alleged copyright infringement. We agree with this assessment and do not see a need for intervention on this issue.

Conclusion

29. The Government has decided to adopt Option 0 and make no changes to existing protection for CGWs. As the use of AI to generate creative content is still in its early stages, the future impacts of this provision are uncertain. It is unclear whether removing it would either promote or discourage innovation and the use of AI for the public good.

30. Therefore, we intend to monitor the impacts of existing protection as AI develops further. This will include monitoring case law in this area, discussions with stakeholders, and through engagement with international partners. To that end, the IPO welcomes any new evidence stakeholders wish to provide in future on the effect of the CGW regime. We do not rule out changes to CGW protection in the future, should it be needed.

Text and Data Mining

Summary

31. Text and data mining (TDM) means using computational techniques to analyse large amounts of information to identify patterns, trends and other useful information. TDM is used for training AI systems, amongst other uses. It also has uses in research, journalism, marketing, business analytics and by cultural heritage organisations. The consultation outlined options in relation to text and data mining, which developers of AI and other copyright users had identified in the call for views.

32. Although factual data, trends and concepts are not protected by copyright, they are often embedded in copyright works. Data mining systems copy works to extract and analyse the data they contain. Unless permitted under licence or an exception, making such copies will constitute copyright infringement.

33. Some rights holders license their works to allow TDM, but others do not. This has financial costs for people using data mining software. To reduce these costs for researchers, an exception to copyright for TDM was introduced in 2014. This is set out in section 29A of the Copyright, Designs and Patents Act 1988. It is limited to non-commercial research, following EU rules which were in place at its creation.

34. Several other countries have introduced copyright exceptions for TDM. These encourage AI development and other services to locate there. Territories with exceptions include the EU, Japan and Singapore. TDM may also be fair use under US law, depending on the facts.

35. The consultation sought views on how to make it easier for people to data mine copyright materials. It did this with a view to supporting AI and wider innovation in the UK, in line with Government priorities on AI, data and innovation.

36. 65 respondents expressed a preference, with a spread of views across the responses. Almost no quantitative evidence was provided on the licensing environment or other questions.

37. Rights holders favoured no change or licensing solutions to help make more material available for TDM. Users of copyright and database material favoured a wider exception. They highlighted the costs of licensing and difficulties in obtaining licences, especially when many rights holders are involved.

38. The Government has decided to introduce a new copyright and database exception which allows TDM for any purpose, along the lines of Option 4. This option is the most supportive of AI and wider innovation.

Options and responses

39. Responses on TDM were received from individuals, academics, lawyers, representative and trade organisations, and businesses involved in AI or with an interest in TDM. We have also considered views expressed at the roundtables held during the consultation period.

40. Many rights holders stated that there was no evidence of a problem, so licensing should continue to be controlled by rights holders. They were against changes to copyright exceptions. They said that commercial uses of TDM should be licensed, especially outside of scientific research. Some music rights holders argued that music cannot and should not be reduced to data for the purposes of mining. While we respect the creativity involved in copyright and databases, there are datasets in all types of work.

41. Users said that exceptions are the best way to reduce the cost of TDM. Licences can price users out of the market, due to the cost of the licence or the complexity of identifying and agreeing licences with many rights holders. Some users highlighted more attractive conditions in other countries for TDM where broad exceptions already exist.

Option 0: make no legal change

42. Some rights holders argued that there is no quantitative evidence of a need for an exception beyond the one for non-commercial research. They concluded that, in absence of such evidence, the rights holders should be left to control access to data through licensing.

Option 1: improve licensing environment for the purposes of TDM

43. The consultation asked specific questions on the licensing environment, but rights holders did not provide quantitative evidence for this. However, some rights holders described licensing arrangements in place in the market. Some rights holders indicated that they had never refused to issue a licence for TDM. They argued that licensing revenue helps to support the investment needed for curation to make quality data available for mining.

44. Users had mixed experiences with TDM licensing. Some had been priced out of licensing for research projects which have restricted funding. Others indicated that licensing was costly and unworkable when mining content from large numbers of individual rights holders.

45. Various respondents suggested ideas for improving licensing. Some do not need Government intervention and could be taken forward by the market. These include a pilot licensing scheme for small AI developers to access scientific and technical material. Collective licensing could be considered where rights holders are represented by collective management organisations (CMOs).

46. Licensing suggestions with limited Government intervention included the creation of a code of practice for TDM or model licensing terms. It was also suggested to consider better education about licensing in general. These ideas could be achieved through partnership between Government, users and rights holders.

47. Some users thought that a compulsory licensing scheme for TDM might be preferable.

Options 2-4: TDM exceptions

48. There were three options in relation to a TDM exception. Option 2 was to extend the existing TDM exception to cover commercial research. Option 3 was to adopt a TDM exception for any purpose, with a rights holder opt-out. Option 4 was to adopt a TDM exception for any purpose, which does not allow rights holders to opt out. All options would apply to database right as well as copyright.

49. As most respondents discussed the options for exceptions together, we have grouped them in this response.

50. Rights holders were broadly against the introduction of new exceptions. They argued that it interfered with their rights to exploit their works. They also said that it could lead to a decline in the creation and curation of datasets for mining, as they could not license them.

51. Some rights holder respondents indicated that a limited new exception might be acceptable. It should still allow licensing for commercial uses. Some referred to the EU’s data mining exception and the UK’s education exceptions as examples of these.

52. Users generally preferred an exception because it removes the need for licensing. Some respondents felt that an opt-out for rights holders might be appropriate, but the detail of this would be critical. Any opt-out would need to be entirely machine readable to make the exception workable. Several also pointed out the less favourable system in the UK compared with other AI-leading nations with exceptions. The legal framework in any particular country, including the UK, plays a part in investment decisions. Some users suggested that investors would choose to support AI and TDM projects in other countries because of the current limited UK framework.

53. Under option 3, we asked specific questions in relation to the new TDM exception in the EU which contains a rights holder opt-out. Many respondents said that some member states have still not implemented the relevant Directive. So, it is too early to say how it is operating. The opt-out is intended to be machine readable and easy to use, and discussions on appropriate methods are still at an early stage. Some rights holders objected to the opt-out all together, as it places the burden on them rather than the user. They said it would be easier to remove their works from the public completely rather than have a machine readable opt-out.

Other data mining issues

54. Respondents to the consultation raised issues related with data mining, including data protection. The uses of data other than under the copyright framework are outside the scope of this consultation.

55. Data ethics and standards were other areas of interest for respondents. Ethical questions can relate to personal data and what data should be used for training AI and other TDM uses. Data ethics are often considered in relation to scientific research. Standards questions concerned formats and types of data, which can vary even in the same type of copyright work, for example. The regulation of data standards and data ethics are not a part of the copyright framework and are outside the scope of this consultation.

56. Some respondents suggested tax incentives for AI developers to reduce costs. This should help them buy licences from rights holders. Taxation policy is outside the scope of this consultation.

57. Some respondents felt that the introduction of an exception under options 3 and 4 would lead to unfair competition in some sectors. Competition issues are generally a matter for the Competition and Markets Authority. However, the economic impact of exceptions on both users and rights holders will be in the impact assessment accompanying any legislation.

Conclusion

58. The Government has decided to introduce a new copyright and database right exception which allows TDM for any purpose. The Government will identify suitable legislation to make the required changes in due course.

59. Introducing an exception which applies to commercial TDM will bring benefits to a wide range of stakeholders in the UK. These include researchers, AI developers, small businesses, cultural heritage institutions, journalists, and engaged citizens. Targeted products and services will benefit businesses and customers. Research outcomes could also benefit the wider public. This could be, for example, by supporting research and innovation in public health. Some in the creative industries also use TDM and AI to understand their market or create new works – they will also see benefits. The benefits will be reducing the time needed to obtain permission from multiple rights holders and no licence fee to pay. This will speed up the TDM process and development of AI.

60. These changes make the most of the greater flexibilities following Brexit. They will help make the UK more competitive as a location for firms doing data mining.

61. Rights holders will no longer be able to charge for UK licences for TDM and will not be able to contract or opt-out of the exception. The new provision may also affect those who have built partial business models around data licensing. However, rights holders will still have safeguards to protect their content. The main safeguard will be the requirement for lawful access. That is, rights holders can choose the platform where they make their works available, including charging for access via subscription or single charge. They will also be able to take measures to ensure the integrity and security of their systems.

62. The Government’s ambition is to make the UK a global centre for AI innovation. The new exception will ensure the UK’s copyright laws are among the most innovation-friendly in the world. All users of data mining technology will benefit, with rights holders having safeguards to protect their content.

Patents

Summary

63. We received 48 written responses that expressed a view on patents and AI inventorship. Respondents included AI developers, the pharmaceutical sector, small businesses, academics, and legal and IP professionals. Respondents had mixed views on the options proposed in the consultation. The majority preferred no unilateral change in UK law for now (Option 0). There were no trends in preferences for any options according to industry sector, business size or business model. The majority of respondents consider AI is an assistive tool. Many stated that inventions developed using AI are protected enough by the current law.

64. The views expressed in the roundtables are in line with the consultation responses. That is, there should be no legal change to UK patent law and any future change would need to be at an international level.

65. We have decided to adopt Option 0: make no changes to the law. There is no evidence that UK patent law is currently inappropriate to protect inventions made using AI. We are also sensitive to concerns that unilateral change now, risks being counterproductive. So, we will advance international discussion so that inventions devised by AI are appropriately protected in the future.

Patents options

66. The proposals for legal change featured in the consultation raised concerns with respondents. The majority of respondents highlighted issues around diverging from international norms on inventorship. They felt a change only in UK law would prejudice international patent filings in markets that are important for UK interests. Many respondents who supported legal change urged the UK to take the lead in this area. However, most respondents agreed that any change to the law on inventorship must be harmonised internationally.

67. Some people said that rules will need to change to take account of the increasing contribution of AI in the research and development process. This view was a particular concern of the pharmaceutical industry. Many respondents felt that once AI reaches a stage where it could invent, any inventions devised by that AI should be patentable. But respondents from the tech sector acknowledged that AI is not yet advanced enough to invent without human intervention.

68. Most respondents share the view that there is limited technological, legal, and economic evidence for legal change at present. In particular:

• there is no consensus on who should own patents when AI is the inventor
• respondents consider the technological case for change may be stronger in the future
• international harmonisation is seen as a prerequisite to successful change
• the majority took the view that a unilateral change to UK legislation would be premature and may be negative for the UK’s innovation landscape

Option 0 – No legal change

69. Most comments state that no legal change is the best option at the moment. Almost all respondents share the view that AI is a tool used by human inventors. Many respondents believe AI is not capable of inventing without significant human intervention. Many respondents believe the current IP framework protects AI-assisted inventions enough.

70. Most respondents agreed that future developments in AI may require legislative changes at an international level to protect inventions devised by AI. There is no clear time scale for technological developments in AI. Some respondents from the pharmaceutical sector felt that the law should stay ahead of technological developments. The UK should be pro-active in this area. Many thought there is a risk that inventions made using AI will not qualify for patent protection in future.

Option 1 – Expand the definition of inventor

71. For those who supported legal change, the majority supported Option 1 over Options 2 and 3. Under Option 1, the definition of inventor would be expanded to allow humans responsible for the AI to be named inventor. Some respondents felt this option would be more palatable at an international level. Others were critical of this option and felt it would be disingenuous to name a human inventor where AI is the sole deviser.

72. There was no consensus on how to do this without detrimental ownership consequences. Many supported applying a test for inventorship that the inventor is the person who made the necessary arrangements for the creation of the invention. Ownership would then be decided in the usual way. Others proposed an adaption of the current employer/employee relationship. Some respondents suggested that contracts should settle issues of patent ownership.

Option 2 – Allow AI to be named as inventor or remove requirements to name an AI inventor

73. Option 2 would allow an applicant to name AI as inventor. A few supported this option but only if legal change were implemented internationally. Many expressed the view that a premature change by the UK would significantly affect an applicant’s global filing strategy. It may lead to a scenario whereby an invention can only gain patent protection in the UK.

74. Representatives of SMEs suggested that Option 2, may have a negative financial impact on small businesses. A small business may incur greater legal costs when seeking advice on global filing strategies and compliance with UK inventorship rules. Patent entitlement, which stems from inventorship, was highlighted as an additional concern for SMEs. Many SMEs worry that a change in the law may allow a third-party AI developer to claim inventorship rights over their invention.

75. The small number advocating for a change in the law argued that Option 2 would encourage innovation and investment in AI technology. Some said that allowing AI to be named as the inventor would incentivise the development of AI and make inventive AI a valuable commodity. Many felt the need for transparency is important where AI has made a significant contribution to an invention. Very few supported the option to remove the obligation to name an inventor.

Option 3 – New patent-like right to protect inventions devised by AI

76. The option to introduce a patent-related right was only supported by few respondents. The consultation did not give a clear picture of how this option should be implemented to best support the UK economy. A small number identify some benefits of introducing a new intellectual property right. Some stated that it may provide the opportunity to create a bespoke, flexible regime for AI devised inventions. Others noted it may provide an opportunity for the UK to better understand AI’s contributions to innovation. A new patent-like right may also give AI technology the chance to develop whilst benefiting from registered IP protection.

77. Most respondents felt introducing a new right for inventions devised by AI would create unnecessary conflict and confusion with the current patent system. Some argued it would fragment an already complex IP landscape.

Other Patent issues

78. We are concerned that the recent publicity about naming an AI system as an inventor risks the incorrect conclusion that UK patent law does not protect AI-assisted inventions. A patent may in fact be granted for an AI-assisted invention provided the application satisfies the legal requirements set out in the UK Patents Act 1977.

Conclusion

79. The Government will make no change to UK patent legislation now. Most respondents agree current UK patent rules on inventorship are enough to protect AI-assisted inventions. Many share the view that any change to rules on inventorship should be harmonised at an international level. We will therefore seek to advance AI inventorship discussions internationally to support UK economic interests and competitive edge. We will need to understand how our IP system should protect AI-devised inventions in the future.

80. The Government does not want to risk creating barriers to innovation and investment in AI technology. We will need to be on the front foot of technological developments in AI. We will take forward the following actions:

• advance AI inventorship harmonisation discussions in international fora
• develop a clear picture of what the UK should seek in harmonisation discussions, through:
  (i) Understanding the existing and potential strengths of the UK economy, and what would reinforce UK competitive advantage; and
  (ii) Actively monitoring technological developments in AI’s use in the innovation process by building on relationships formed during the consultation
• seek to address any perceptions that the current law prevents the patenting of AI-assisted inventions

Other issues raised by respondents

81. Many feel there needs to be a broader ethical and moral debate on whether AI should be recognised as a legal personality. This is not a problem which can be resolved in IP law and falls outside the scope of this consultation.

82. Some respondents raised concerns about the impact of AI on performers. This includes concerns that computer-generated performances might replace human performers. The effect on audio performances (such as audio-book narration and voice-overs) was a particular worry. There are also concerns that computer-generated performances provide greater opportunities for using performers’ images, voices or likenesses without permission. For example, in ‘deepfakes’. Some stakeholders called for an expansion of the scope of performers’ rights in the Copyright, Designs and Patents Act 1988 to address these issues.

83. We take these views seriously. But, at this stage, the impacts of AI technologies on performers remain unclear. It is also unclear whether and how existing law (both in the IP framework and beyond it) is insufficient to address any issues. If intervention is necessary, the IP framework may not be the best vehicle for this. We will keep these issues under review from an IP perspective.

Overall conclusion

84. The Government considered three specific areas:

• copyright protection for computer-generated works (CGWs) without a human author

• licensing or exceptions to copyright for text and data mining (TDM), which is often significant in AI use and development

• patent protection for AI-devised inventions

85. For computer-generated works, we plan no changes to the law. There is no evidence at present that protection for CGWs is harmful, and the use of AI is still in its early stages. As such, a proper evaluation of the options is not possible, and any changes could have unintended consequences. We will keep the law under review and could amend, replace or remove protection in future if the evidence supports it.

86. For text and data mining, we plan to introduce a new copyright and database exception which allows TDM for any purpose. Rights holders will still have safeguards to protect their content, including a requirement for lawful access.

87. For AI-devised inventions, we plan no change to UK patent law. But we will keep this area of law under review to ensure that the UK patent system supports AI innovation and the use of AI in the UK. We will seek to advance AI inventorship discussions internationally to support UK economic interests.