The intersection between AI and Intellectual Property (IP) policy is a complex one and has many implications for key stakeholders both now and into the future. This week I look at the results of a public consultation driven by the UK Intellectual Property Office (IPO) on the matter. Responses released looked at many issues around AI in relation to patents, trademarks, designs, trade secrets, and copyright.

Here, I share insights into some of these responses and specifically explore where manufacturers will most likely be impacted by key developments. Having worked in AI across multiple sectors for several years, I know that the area of AI patent infringement has major challenges. It’s one area that requires robust policy input to protect any manufacturers who make or use AI tech.

The consultation

The call issued 45 questions covering five key areas, that were answered by a total of 92 responders. Submissions came from a range of entities including IP attorneys, trade bodies, industry associations, tech sectors, creative industries, and other sectors. An interesting summary of the questions asked can be viewed here followed by individual responses submitted here.

The five areas covered the legal context and policy background but with a focus on how AI impacts IP. They include;

1. Patents
2. Copyright and related rights
3. Designs
4. Trademarks
5. Trade secrets

There were also questions about IP rights and how AI can help stop counterfeiting and piracy violations. It’s being considered in various areas to help support the enforcement of IP rights. The call looked at the legal framework relating to IP rights but did not consider the use of AI tools to combat IP infringement, which is perhaps one of its shortcomings. It did, however, recognise and support the use of such technology by IP rightsholders.

What does it mean for manufacturing?

I was especially interested in the area of liability around patent infringement, for example, if AI technology contributed to a patent infringement, who should be held responsible? Some of the responses indicated that the AI developer or the manufacturer who has made money from the sale of the AI is accountable. Others believe that the solution lies in whoever developed the AI in the first place.

On a more practical level, could be difficult given that there are often many individuals and teams involved in developing AI systems. There can be many actors involved ranging from those who are inputting and training data to those who are orchestrating the AI’s output.

The technology itself is also becoming increasingly independent of its human developers, so accountability when AI infringes a patent, particularly when this action could not have been predicted by a human, is far from straightforward.

The responses referred to UK law that defines the activities under the control of a patent owner and those carried out without the consent of the patent owner which will infringe a patent. The law only recognises “a person” as infringing a patent. It doesn’t set out how liability works when a person is not involved. In a nutshell, the law assumes that those infringing on the patent know that they have done so, it’s only an infringement if the infringer knows what they have done.

Accountability

Holding humans accountable for an infringement by an AI system is complicated even further if AI systems become “autonomous”.  The adoption of autonomous AI will increase across many sectors but especially in industrial settings.

There is a need for a strong policy framework to protect manufacturers from being liable for patent infringement committed by an AI. For me, this starts educating companies on the legal and ethical implications of AI to their business. This is further compounded as R&D investment in AI increases and AI becomes more commonplace. Market intelligence company IDC forecast European AI spending of $10 billion for 2020 alone, however, the global pandemic altered this and some sectors were more resilient than others. There is a lot to play for!

Finally, it’s not just the UK IPO engaging the public on such issues, last year the World Intellectual Property Organisation (WIPO) published a revised Issues Paper based on hundreds of submissions. This publication looked at many of the same IP policy issues and the full paper can be read below.

Further reading

• Consultation outcome Artificial Intelligence call for views: patents, UK Intellectual Property Office

• Consultation outcome download, UK Intellectual Property Office

• Consultation questions, UK Intellectual Property Office

• Spending in Artificial Intelligence, IDC

• WIPO conversation on intellectual property and artificial intelligence WIPO