IP challenges in the AI space


Financier Worldwide Magazine

January 2018 Issue

Artificial intelligence (AI) has emerged as a powerful force in global business. The potential size and value of the AI industry is enormous, and while the market is still small, it is growing. According to Statista, the AI sector was worth around $2.42bn in 2017; by 2022 it is expected to be worth in excess of $16bn, according to Research and Markets.

AI has the potential to revolutionise many aspects of the working world. It is forecast to improve labour productivity across the globe by around 40 percent and double world economic growth by 2025. Further, according to IDC, worldwide AI revenue should see a compound annual growth rate of around 57 percent to $47bn between 2016 and 2025. This growth will be generated by improving productivity, cutting costs and providing a competitive advantage.

AI will not only disrupt business models, but also legal frameworks. The creation and utilisation of AI will undoubtedly give rise to new intellectual property (IP) challenges, including the use of AI to create IP, to carry out administrative tasks, and to manage IP regulation, among others.

AI creations

Given its increasing sophistication, human creators are utilising AI to co-create works, or in some cases, almost entirely generate content itself. Machine learning – a subset of AI that produces autonomous systems that are capable of learning without being specifically programmed by a human – is being harnessed to generate a variety of IP. These programmes have built-in algorithms which allow the system to learn from datasets and produce increasingly sophisticated work. But this process can be problematic. Ownership of the IP in the dataset used to train the system must be considered. Including data without permission presents a potential liability risk which could hinder the development and commercialisation of the system in the future.

In addition, depending on the jurisdiction in which the IP is created, the ‘author’ of the work is unclear when AI has acted as the sole creator. UK legislation states that where a work “is generated by a computer in circumstances such that there is no human author of the work” the author is “the person by whom the arrangements necessary for the creation of the work are undertaken”. This does not allow the AI itself to be considered the author and could lead to disputes, particularly over whether the inputs are strong enough for the person who provided them to be classed as the ‘author’. Furthermore, issues abound over who made the arrangements. Should the author be the person that built the system, the person who trained it, or the person who fed it these specific inputs? For affected organisations, it would be advisable for them to clearly set out who is going to own the IP in any commercial agreements and terms of use for the system.

Existing legal frameworks are not ready to confront many of the issues created by the rise of AI.

Existing legal frameworks are not ready to confront many of the issues created by the rise of AI. Legally, AI will muddy the waters of IP ownership, particularly if several parties were involved in development. As AI becomes more sophisticated in the coming decades, a significant shift in these frameworks will be required. Liability will also be an issue.

Much like the UK, many other jurisdictions are unwilling to consider AI-generated works eligible for copyright. In Spain and Germany, copyright legislation states that only work created by a human can be protected by copyright. In the US too, the Copyright Office has declared that it will “register an original work of authorship, provided that the work was created by a human being”. These legislative stances will likely change in the future, as AI becomes more sophisticated and the line of inventorship becomes more complicated as a result.

AI has the potential to generate real value for companies. It can be harnessed to not only generate IP content but to undertake mundane, day-to-day IP management tasks, which will have cost and time implications. Automation of these processes will allow companies to focus their energy on more strategic issues. AI will allow IP professionals to more quickly and accurately value their IP portfolios and better understand where they should be investing.

Research published by the University of Surrey in the Boston College Law Review argues that existing legal frameworks have failed to address the issue of AI invention. Given the forecast increase in the volume of AI generated IP in the coming years, it would be prudent to acknowledge computers as inventors in order to incentivise the development of creative computers.

Regardless of the pace of legislative change, AI will help companies open up new markets and create opportunities. However, legislative changes will be required to fully embrace the potential offered by AI in the IP space and to clarify uncertainties surrounding AI-generated IP.

© Financier Worldwide


Richard Summerfield

©2001-2018 Financier Worldwide Ltd. All rights reserved.