The use of artificial intelligence (AI) and machine learning is growing at a significant pace and spreading across many industry sectors, including healthcare. With the rapid development of AI technology which has the potential to revolutionise many aspects of our lives, including in providing and receiving healthcare services, the concept of “creations of the mind” is no longer limited to creations by a human being. These technological developments mean that the legal framework governing intellectual property (IP) rights such as patents and copyright, which protect “creations of the mind”, may need to be adjusted to address the changes and impacts brought about by the use of AI.
In line with the UK government’s ambition for the UK to be a leader in AI and to better understand the implications AI might have for IP policy, as well as the impact IP might have for AI in the short to medium term, the UK IPO conducted a public consultation at the end of 2020. The aim of the consultation was to seek responses on a range of questions relating to AI and IP rights. The UK IPO received 92 responses from a wide range of stakeholders, including IP rights holders, producers of AI technology and academia. The government’s response to the call for views on AI and IP was published in March 2021, under which reforms to patent and copyright law and policy were discussed.
In this blog, we summarise the UK government’s conclusions from the consultation before considering the potential impact to digital health applications and companies.
Potential changes to UK patent and copyright law and practice
The UK government’s response to the consultation indicates that it will consult later this year on a range of possible policy options, including potential legislative changes to the Patents Act 1977, to protect AI generated inventions which would otherwise not meet inventorship criteria. The current position in the UK, as confirmed by the High Court’s decision in September 2020 in the case of the “DABUS” AI machine, is that the applicant of a patent application must be a “person” with legal personality. Legislative changes to the inventorship criteria could make inventions derived from AI machines patentable.
The UK government’s response also proposes other measures to address suggestions and feedback received in the consultation relating to the conditions for grant of a patent for AI inventions, including:
- publishing enhanced guidelines on patent exclusion practice for AI inventions;
- commissioning an economic study to enhance the understanding of the role of the IP framework in incentivising investment in AI; and
- considering establishing a deposit system for data used to train AI systems disclosed within patent applications.
The UK government intends to review and consult on the AI licensing framework, as well as the scope of copyright exceptions with a view to supporting innovation and research in AI. Exceptions to copyright infringement include, in certain cases, copying to enable technology to work more effective (for example, temporary copies to be made during web browsing and signal processing), as long as these copies do not have independent economic significance and enable a lawful end use. Another exception to copyright infringement is text and data mining (TDM) for non-commercial research. However, given that the majority of AI tools and processes are developed for commercial processes, this TDM exception is of limited applicability.
The UK government also intends to consult on whether to limit copyright in original works to human creations (including AI-assisted creations) and whether it should replace the existing protection for computer-generated works with a related right.
No imminent changes to trade mark or design right laws are expected from this consultation exercise, nor for trade secrets protection. Nevertheless, given the rapid pace of adoption of AI across many sectors, the UK government has said that it will keep the impact of AI on the IP framework under close review.
Potential implications for the digital health sector
On 16 June 2021, the UK government announced 38 new pioneering AI projects with a total award value of £36 million to help the NHS to transform the quality of care and the speed of diagnoses. The AI Roadmap independent report published by the AI Council in January 2021 also lists “using AI to improve outcomes and create value in healthcare” as one of the 16 recommendations to help the government develop a UK National AI Strategy.
The potential reforms to UK patent and copyright law and practice could further benefit the digital health sector. If legislative changes come into effect in the UK to remove the current requirement of legal personality for inventorship, digital health inventions made in whole or in substantial part by AI systems could potentially be protected by patents.
As pointed out by the UK BioIndustry Association (BIA) in its response to the consultation, the key question for inventorship is to determine, based on the facts of each invention, who is the “actual deviser” of the invention. As AI technology develops, it is conceivable that in some situations in the future, the AI system could be the “actual deviser” of an invention with no or minimal human being involvement, and in that case, it would be appropriate to allow the AI system to be identified as the inventor. The problems with not being able to credit AI as an inventor in such cases, as pointed out by the Chartered Instituted of Patent Attorneys (CIPA) in its response, could be that:
- a person may be identified as the “deviser of an invention” just to provide a basis for a patent application, even though there may be uncertainty over whether his or her contribution satisfies the legal requirement for inventorship. This may create problems downstream, such as issues over the ownership and validity of the patent;
- a proprietor may downplay in a patent application the contributions made by an AI system, thereby reducing the accuracy and value of the information provided in the disclosures with the application; and
- it may discourage further research and development using an AI system because the research results and innovation are not patentable.
Other responses highlighted potential risks in expanding protection of AI related inventions. For example, the UK chapter of the International Association of the Protection of Intellectual Property (AIPPI UK) noted in its response that allowing an AI system to be named as an inventor may motivate a “land grab” on technologies by a few companies that have the financial resources to deploy AI and they warn of a patent “arms race” between competitors. Further, AIPPI UK pointed out that the patent system exists to reward “the spark of human invention in return for disseminating knowledge to human society”, instead of the automatic generation of ideas (however inventive) by an automated process with no human intervention.
Another potential benefit to the digital health sector is the increased clarity as to patentability of inventions that use AI and machine learning. Although digital health inventions that use algorithms, machine learning, data analytics and computational methods are excluded from patentability under both the UK Patents Act 1977 and the European Patent Convention, this is only to the extent that an invention only relates to such excluded subject matter. The UK and the EPO have developed differing legal tests when assessing such patent exclusions. The responses to the consultation included the common view that the UK IPO should change its practice when assessing patent exclusions, rather than amending the underlying legislation. As the UK and the EPO share the same categories of excluded subject matter, some of the commentators (such as the BIA) noted in their response that it is important to avoid divergence between the UK and the EPO in dealing with patent exclusions as this creates an extra burden for businesses and the UK IPO should follow the patent exclusion practice of the EPO.
The government’s response to this issue acknowledges that UK patent exclusion practice needs to offer more clarity to improve predictability in relation to AI implemented inventions. Rather than consult on changes to the underlying legislation, the UK government’s proposed next steps focus on the publication of enhanced IPO guidelines on patent exclusion practice for AI inventions and the engagement of AI interested sectors and the patent attorney profession to enhance understanding of UK patent practice.
The possibility of a new right to replace existing protection for computer generated works, with a scope and duration reflecting investment in such works would provide an additional form of protection for digital health businesses. This may be analogous to the sui generis database right, which also incentivises investment rather than creativity. On the other hand, the proposed consultation to limit copyright in original works to ensure that works generated solely by AI are not protected, might restrict the copyright protection afforded to digital health businesses that use AI systems to generate data.
TDM is increasingly being used in health research and to develop health applications. As the TDM exception to copyright infringement does not apply to commercial activities, digital health companies must seek licences for TDM from the rights holders, but contractual restrictions in such licences can limit access and the potential value of the data sets. The UK government is taking a prudent approach on the issue of whether the TDM exception should be expanded by first furthering its understanding of the merits of the TDM exception, including drawing on the approaches taken in other countries. Should it decide to make legislative changes, it will need to strike a careful balance between the need to protect the interests of rights holders and the need to meet the demand for access to works by technology firms, entrepreneurs, and researchers to promote innovation.