UKIPO Consults Further on AI and Copyright and Patent Inventorship

The UK Intellectual Property Office (UKIPO) has issued a second consultation on artificial intelligence and intellectual property, this time focussing on what changes are required for copyright and patents laws in order to facilitate the UK being a leader in AI research, development and implementation.

The earlier consultation sought views on how AI impacts on and is affected by all facets of IP. While potential issues in relation to trade marks, designs and trade secrets were raised during the consultation process it was considered that there is no urgent need for change as the law in these areas continues to be fit for purpose. However, it was acknowledged that the impact of the developing use of AI on those areas and the extent to which those areas constrain the development of AI should be monitored. A more detailed review of copyright and patent law, though, is considered to be more pertinent and pressing in view of the issues identified through the earlier consultation.

Along with most other forms of IP, copyright and patent rights are conditional incentives based mechanisms designed to encourage the creation and disclosure of original, inventive or distinctive works or ideas. However, there are notable respects in which the creation of copyright or patentable subject matter by or through the use of AI is not incentivised or is hindered by the current law. For instance AI is very data intensive and its outputs can be affected if copyright works are excluded from its inputs. Doubts about the ability to name AI as an inventor or co-inventor can create doubts about how to validly apply for a patent when an invention is created by or with AI.

The current consultation focuses on the following three specific areas:

1. 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 should they be protected?
2. Licensing or exceptions to copyright for text and data mining, which is often significant in AI use and development.
3. Patent protection for AI-devised inventions. Should we protect them, and if so, how should they be protected?

Further focus is provided by specifying that to be implemented any proposed changes should:

1. Encourage innovation in AI technology and promote its use for the public good;
2. Preserve the central role of intellectual property in promoting human creativity and innovation;
3. Be based on the best available economic evidence.

Computer-Generated Works
The UK is one of only several countries, including New Zealand and Hong Kong, that provide copyright protection for computer-generated works. Some argue that computers do not need to be incentivised to create new content and that the existence of such protection creates costs for third parties. Given that computer-generated works have a shorter period of protection than human authored works some also argue that it encourages false attribution of a computer-generated work as being authored by a human, although the UK Fraud Act 2006 is considered to adequately discourage such claims.

There are three broad options for copyright in computer-generated works: (i) make no change; (ii) remove protection for such works; and (iii) replace the current right with a right having reduced scope and duration. To justify the status quo option it would need to be shown that the generation of new AI works is encouraged. The removal of protection would not affect human authored works created with the assistance of AI. To justify the removal option it would need to be shown that copyright protection is not necessary to incentivise the production of computer-generated works or that unreasonable costs are created for third parties. To justify a new right it would need to be shown that the generation of new AI works is encouraged if there is a shorter term of protection and scope of rights.

Text and Data Mining
Currently there is a conditional exception that allows for the use of copyright material for non-commercial text and data mining research purposes. Research shows that this is being used as intended, although may benefit from updated guidance. Other options are to: improve the licencing environment through the provision of model licences or codes of practice; extending the exception to include lawfully accessed commercial uses with attempts at contractual override being unenforceable; and whether or not to allow rights holders to opt-out particular works or collections of works.

Inventorship for Patents
As recently confirmed by the Court of Appeal (although possibly subject to an appeal to the Supreme Court) the UK restricts patent inventorship to natural persons and limits claims to inventorship to the actual devisors of the invention. This status quo option allows patents to be granted where AI has been used as a tool by a human inventor, but not if AI is named as the or a devisor of the invention. Other than the status quo the other options for consideration are:

• to widen the definition of inventor to include the humans responsible for the AI system that devises an invention. This could allow the people with the following responsibilities to be considered inventors: programming the AI, configuring the AI, operating the AI, selecting input data such as training data for the AI or recognising applications of the output of the AI.
• allowing AI to be named as the or an inventor by either specifically allowing AI to be an inventor or by removing the requirement that the inventor is a human. Until other jurisdictions also allowed this obtaining patent protection in other jurisdictions would be problematic.
• allowing protection for AI devised inventions through a new right with a shorter term that may either have a stricter test for obviousness or no test for obviousness.

Author: Quinn Miller

//piperpat.com/news/article/ukipo-consults-further-on-ai-and-copyright-and-patent-inventorship