December 01, 2021

UK Government's consultation on artificial intelligence and the interaction with copyright and patents


Related Author:   Kahroba Kojouri, Trainee Solicitor at Mayer Brown

The UK Government published its national AI Strategy last month which highlighted its intention to make the UK an AI powerhouse. In its strategy, the Government recognised the importance of instituting a progressive regulatory environment which would encourage the advancement of AI. The UK Government has now launched a consultation on reforms to copyright and patent law to incentivise the development of AI. Whilst the Government aims to provide adequate protections to AI technology to encourage further investment into AI, it is cognisant that those protections have to be limited so that they do not stifle innovation and progress in this field. The UK Government proposes that any measures introduced must (i) encourage AI innovation and promote the use of AI for the public good; (ii) preserve the integral role of intellectual property, which is to promote human creativity and innovation; and (iii) be based on best available economic evidence.

There are three specific areas which the Government is consulting on:

1. Copyright protection for computer-generated works without a human author

The UK currently protects works generated by a computer without a human creator by granting copyright in the works to the person who produced the computer program which generated the works. The protection lasts for 50 years starting from the date on which the works are created. Critics argue that for a work to benefit from copyright it must be original; however, the legal concept of originality is defined in the law by reference to human authors and characteristics, such as personality, judgement and skill. Therefore, the law is contradicting itself by providing copyright protection to computer-generated works without a human author. Further, they argue that providing such works protections may promote them at the expense of human creations, ultimately devaluing human creativity. They also contend that the protection granted is excessive since it does not actually reward its maker, but comes at a tangible cost for third parties. The Government proposes three options with respect to computer-generated works:

  1. Make no change to the law so that computer-generated works continue to benefit from copyright protection, especially if it encourages the development of new AI.
  2. Remove the copyright protection that computer-generated works currently benefit from so that copyright is limited to works made by human creators. AI works that require a sufficient level of human intellectual creativity or meet the originality requirement may continue to benefit from copyright protection. This option would be justified where copyright protection for computer-generated works is not needed to incentivise the production of new AI.
  3. Replace the existing right with a new right which provides a reduced duration of protection (for instance, five years). It would sit alongside any other rights that subsist in the works, as such, works can benefit from this protection where they have been produced via a co-creation by a human and an AI system. Again, this option would be justified if the evidence suggests that limited protection for computer-generated works incentivises the production of AI and presents a better balance between the interests of right holders and those of third parties.

2. Licensing or exceptions to copyright for text and data mining

Text and data mining (TDM) consists of the use of automated computational techniques to analyse large amounts of information to identify patterns and predict trends. TDM can be useful to both develop and train AI. Some of the data captured to be analysed in TDM can be protected by copyright; therefore, a copyright license to use the data needs to be obtained or the exception provided under the Copyrights, Designs and Patents Act (the "Act") needs to be relied upon. The exception in the Act allows TDM copying where: (i) copies are made for non-commercial research; (ii) the access to the research is lawful (via subscription or permission by way of terms and conditions); and (iii) the copy is accompanied by an acknowledgement of the works and its right holders (where practical). The exception cannot be overridden in contracts and publishers of such content may also apply reasonable measures to maintain their network security or stability. Further, this exception does not apply to database rights; therefore, all TDM of databases would require obtaining a license.

The Consultation outlines five options to take with respect to TDM:

  1. Make no changes to the exception provided under the Act, but produce additional guidance which provides greater clarity about when the exception can be used.
  2. Provide educational materials, produce model licenses and adopt codes of practice that will improve the licensing environment so that businesses are encouraged to enter into licensing arrangements for use of TDM.
  3. Extend the exception under the Act to cover commercial scientific research and databases where such access is lawful and allows publishers to maintain reasonable measures for network security and stability.
  4. Adopt a TDM exception to works protected by copyright and database rights for use by anyone for any commercial and non-commercial reason, where users have obtained lawful access to the works. Copies obtained can be retained for as long as necessary under this option; however, individual works or sets of works may be opted out of mining through a machine-readable opt-out. This option, for instance, reduces the high costs associated with mining works available on the open internet where it will be difficult to track copyright holders and enter into licenses with them.
  5. Adopt a TDM exception for use of any copyright work or database for commercial and non-commercial reasons, without permitting any works to be opted out of such use. Users would still require lawful access to the works under this option.

3. Patent protection for AI-devised inventions.

Under the current law, an invention can obtain patent protection only where a human is identified as the deviser of the invention. Typically, the individual responsible for creating the AI that then devises an invention is identified as the deviser of the invention. The Government is concerned that the current position may lead to uncertain claims of human inventorship. The Government wants to ensure that adequate protections are given to AI-devised inventions to encourage investment in the field of technology; however, it recognises that an increase in patents granted to AI invented works may have a detrimental effect on innovation and competition, since it is possible that the majority of the patents will be held by a few dominant players who will have access to the best AI technology, potentially undermining the work of start-ups and small and medium sized businesses. 

The Consultation proposes four options to deal with AI-devised inventions:

  1. Make no changes to the legal framework surrounding patents, which would mean that AI cannot be named as a deviser of an invention, but a natural person can continue to be identified as an inventor of the invention where they have used AI tools to devise the invention. This would be advantageous for international businesses since it is consistent with the framework adopted by other countries and would facilitate foreign patent filings.
  2. Expand the definition of "inventor" to include the individuals responsible for programming, configuring or operating the AI system which devises the invention. However, the work undertaken by the individual must have been necessary to the creation of the AI system and mere involvement would not be sufficient. This resembles the first option in that the courts would take a similar approach in identifying the "actual deviser". This would address the issue of obtaining patents for inventions that were entirely devised by AI and may encourage investment into research and development.
  3. Allow patents to be granted to inventions devised by AI by amending legislation to either allow AI to be named as inventors in patent applications or remove the requirement for an inventor to be named in a patent application where the invention has been devised by AI. In both instances, the individual closely responsible for the invention would own the patent and not the AI. Critics of this approach suggest it might be challenging for businesses to gain patent protections in other jurisdictions that only recognise human inventors when their UK patent lists AI as the inventor of the invention, as such, this will discourage applicants from declaring AI as inventors in their patent applications.
  4. Create a separate type of protection for AI-devised inventions where a human inventor cannot be identified. It will operate alongside the patent system in place. The duration of protection would be reduced in this instance since AI can devise inventions much faster and more efficiently than humans. Further, the government suggests that a new stricter test of inventiveness can be introduced to benefit from this protection. Alternatively, it proposes that new AI-devised inventions could be automatically protected and the courts can then decide on the validity of the protection.

The UK is not alone in acknowledging the profound impact AI will have on businesses worldwide. China has also recently declared its intention to become “a principal world centre of artificial intelligence innovation” by 2030. Other nations that are entering the race to dominate the AI space include Germany, the United States and Japan; countries that are currently leading in AI research. Implementing legal frameworks that encourage AI development and investment will play a fundamental role in achieving dominance in this field, which is precisely why the UK Government has published this consultation. Whilst different nations have taken different approaches to the protection of AI created works or inventions (for instance, see the converging views courts have taken with respect to identifying AI as an inventor of a patent in the DABUS cases), there has been a drive amongst states and international bodies to collaborate on laws and regulations governing AI to maximize advancement for mutual benefit. In fact, international cooperation is an element in the AI strategies of most governments. Whether a consensus can be reached amongst nation states with respect to legal frameworks surrounding AI protection remains to be seen.

Organisations are welcome to respond to the UK Government's consultation until 7 January 2022. Details on how to submit a response to the consultation can be found here.


Related Services & Industries

Stay Up To Date With Our Insights

See how we use a multidisciplinary, integrated approach to meet our clients' needs.