AI and IP – A real conversation starter

Countries and organisations around the world are paying greater attention to the intersection of artificial intelligence (AI) and intellectual property (IP).  The UK Intellectual Property Office (UK IPO) recently put out a call for views on the implications of AI for IP policy and conversely, the impact of IP on AI.  The consultation covers five key areas of IP – patents, copyright, designs, trade marks and trade secrets.

Patents

The UK IPO is keen to understand the role of the patent system in encouraging use and development of AI.  A key question is whether patent law should allow AI to be identified as the sole or joint inventor, and if not, whether this will discourage future inventions being protected by patent (and instead, being kept confidential).  Input is sought on the morality and efficacy of this approach and on the level of human input into AI systems that create inventions.

Also open for discussion is the legal framework applicable to AI inventions and liability in the event of infringement.  For instance, the traditional principles behind patents (being exclusive rights granted on a temporary basis in exchange for making the details of an invention publicly available), and the usual grounds for establishing infringement of a patent, may not apply where an AI retains these details in a ‘black box’.

Copyright and related rights

The call for views raises questions on three key areas relating to copyright and AI:

  • the use of copyright works and data by AI systems;
  • whether copyright exists in works created by AI, and who it belongs to; and
  • copyright protection for AI software.

The UK IPO is seeking inputs on whether the copyright framework should make it easier for AI to use protected content – either by introducing new exceptions (e.g. for training AI) or increased facilitation of licensing.

The call for views points out that AI is already used to create copyright works, such as music and art. There are conflicting views on whether content generated by AI should be eligible for copyright protection, on what basis and who should benefit from it.

Inputs are also requested as to whether there is a need for greater clarity about who is liable when an AI infringes copyright and whether software that incorporates AI is afforded sufficient protection by copyright.  We have written more about this topic here: https://www.kwm.com/en/au/knowledge/downloads/kwm-on-ai-artificial-intelligence-ownership-20200227

Designs

According to the UK IPO, AI cannot currently be recognised as the author or owner of a design because AI does not have legal personality.  Views are sought on the need for reform in this regard and in relation to AI’s ability to infringe designs and the attribution of liability should infringement occur.

Trade marks

UK IPO’s primary focus in respect of trade marks is infringement.  The traditional principles relating to trade mark infringement are predicated on human interaction with branding and purchasing decisions.  The increasing involvement of AI in these processes has implications for the concept of the “average consumer” in assessing likelihood of confusion and the imposition of liability in the event of infringement.

Trade secrets

Entities may choose to protect certain IP by keeping it secret or may be able to rely on the trade secrets regime where protection via other IP rights is unavailable or inappropriate.  However, an AI development that is protected as a trade secret (rather than as a patent, for instance) will not be made publicly available.  The UK IPO is seeking views as to the impact this may have on innovation and ethical oversight of AI.

Parties wishing to submit views can do so by email to [email protected].  Consultation closes at 11:45pm on 30 November 2020.

Everyone is talking about AI

The UK IPO emphasises in the call for views that the UK government’s objective is to make the UK a global centre for AI and data-driven innovation, and to increase the uptake of AI for the benefit of everyone.

The ever-increasing use and fundamental impact of AI means we are at a critical cross-roads where IP laws must keep pace or risk not being fit for purpose, or worse, inhibiting the digital future.

In September 2019, the World Intellectual Property Organisation (WIPO) launched a Conversation on IP and AI between its 193 member states and other stakeholders to identify key issues for policymakers.  As part of the conversation, WIPO has released a number of publications, hosted several discussion sessions and produced an informative virtual experience that is available until 18 December 2020.  The next discussion session is due to take place in November 2020.

The US Patent and Trademark Office (USPTO) launched its own public consultation process in October of 2019 to gather information on the impact of AI on IP law and policy.

The Australian government has recognised the importance of building a digital economy, digital transformation in the public sector and investment in science and technology, as seen most recently in the 2020 federal budget. It is also conscious of the many issues raised by the increasing prevalence of AI.  The government is also developing an AI Ethics Framework and has already published a set of voluntary AI Ethics Principles following a public consultation process.

IP Australia has released various publications relating to AI, including in its 2020 IP Report. IP Australia has also implemented AI and automation as part of its own systems and processes. In 2019 it launched IP GAIN which operates like a market place that third parties (including other IP offices) can use to access IP-related AI tools, to improve international collaboration and co-development.

This article was prepared with the helpful assistance of Tom Gilbert.