What’s the impact of Artificial Intelligence on Intellectual Property?
Artificial intelligence is set to revolutionise innovation processes in the coming months and years. Already, it has delivered huge leaps forward in areas such as biotechnology, medicine and chemical research. As an article in Nature describes: “Pfizer is using IBM Watson, a system that uses machine learning, to power its search for immuno-oncology drugs. Sanofi has signed a deal to use UK start-up Exscientia’s artificial-intelligence (AI) platform to hunt for metabolic-disease therapies, and Roche subsidiary Genentech is using an AI system from GNS Healthcare in Cambridge, Massachusetts, to help drive the multinational company’s search for cancer treatments.”
The rise of DABUS
However, its use is ever expanding far beyond healthcare. Most recently, AI hit the mainstream headlines again, when it was reported that a group of academics and an inventor had teamed up and filed for patents on technical inventions on behalf of the computer algorithm that created them, while also making the call for AI to be recognised as the rightful inventor.[1]
The machine in question in this example is DABUS (short for ‘Device for the Autonomous Bootstrapping of Unified Sentience”) — and it has already gained notoriety in the intellectual property world for testing the concept of copyright, due to its ability to independently create abstract art. As an article in WIPO Magazine, published at the time, explained: “Traditionally, the ownership of copyright in computer-generated works was not in question because the program was merely a tool that supported the creative process […] But with the latest types of artificial intelligence, the computer program is no longer a tool; it actually makes many of the decisions involved in the creative process without human intervention.”[2]
DABUS was built by Stephen Thaler, Ph.D, the culmination of many years’ research and the outcome of numerous patent waves that finally led to this technology.[3] In addition to music and images, DABUS, or ‘The Creativity Machine’ has more recently come up with two inventions: a type of food container, optimised for robot picking and packing[4] and a type of warning light that is harder for humans to ignore than conventional beacons.[5] Both of these are creations of technology — with no human intervention — so should DABUS therefore be named as the inventor?
Patentability and AI
The UK patent office, one of the offices that has received these applications, has indicated that the inventions may indeed appear to satisfy the conditions for patentability. Meanwhile, the group, led by Professor Ryan Abbott at the University of Surrey, is looking to pursue the cases further in order to force patent offices around the world to consider the implications of granting a patent to an invention that has not been created by a human.
Professor Abbott is appealing for funds on CrowdJustice, an online crowdfunding platform specifically designed to raise funds for legal action, in order to continue prosecuting the applications. As Professor Abbott explains on the crowd funding website: “There is a near-complete absence of law on the subject of AI-generated inventions. This is a problem, because in the absence of such laws, it is not clear if AI-generated inventions could receive patent protections, who, or what, would be the inventor, and who would be the owner of any inventions.”[6]
Despite the media attention, patent offices are reluctant to consider a knee-jerk reaction to the challenge. As a spokeswoman for the European Patent Office explains: “The current state of technological development suggests that, for the foreseeable future, AI is a tool used by a human inventor. Any change… [would] have implications reaching far beyond patent law, i.e. to authors’ rights under copyright laws, civil liability and data protection.”[7]
Calls for comment
In the United States, the USPTO, which has also received patent applications on behalf of DABUS, recognises that no rapid changes should be undertaken without long and detailed conversations with the wide range of stakeholders that would be impacted by any revisions to patent law. As such, it has issued a request for comment from the public, stating: “The United States Patent and Trademark Office (USPTO) is interested in gathering information on patent-related issues regarding artificial intelligence inventions for purposes of evaluating whether further examination guidance is needed to promote the reliability and predictability of patenting artificial intelligence inventions.” It continues: “The questions are designed to cover a variety of topics from patent examination policy to whether new forms of intellectual property protection are needed.”[8] Those interested in contributing have until 11 October 2019 to provide comments.
In order to provide comment on such matters, there are a number of considerations that need to be assessed as part of this conversation, including the extent to which AI challenges the philosophical, political and economic function of intellectual property. Intellectual property’s purpose is to protect and promote innovation, while preventing the exploitation of one group in relation to another. Imagine, then, that it is accepted that a machine can be named as inventor. There are some immediate questions:
· In a court case, what happens when the inventor is required to give evidence — who could be called upon in this instance?
· During the discovery phase of litigation, what should happen here with respect to the software and source code related to the AI machine?
· Another key phase in a patent litigation is the deposition phase, which involves the ability to quiz witnesses and experts. How does one depose an AI inventor?
· Patents are controlled by international agreements — how easy would it be to reach a new agreement whereby all parties agree that AI should be recognised as the inventor?
· How should regional differences be taken into account, for example in terms of inventor reward and remuneration, which is a key matter in countries like Germany?[9]
· If there are errors in the software that lead to claims, who is accountable?
· Are we likely to treat AI in the same manner as a human inventor? Unless we are going to treat AI as if it were on an equal footing with a human, should it not be that AI remains a tool of the inventor, as the patent offices have so far maintained?
Inventorship and Ownership
Part two of the discussion would then need to focus on the following: If we accept that AI represents inventorship, what does this mean for ownership? Who owns the creation in this scenario? There are at least three possibilities to consider:
· AI is the inventor, therefore the machine owns the creation
· The owner of the machine is the owner of that machine’s inventions
· The data used in the machine is likely to be derived from numerous data sources, including potentially both public and private data. As such, the algorithm would be using information that does not belong either to the machine, or entirely to the machine’s inventor. Arguably, therefore, the output belongs to at the very least the data providers as well as the machine’s owner or, in its broadest interpretation, society as a whole.
It is most likely that the owners of AI invention machines will see option two in this list as the most favourable. Indeed, as an article in Gizmodo, covering the DABUS case (and quoting the researchers) shows, this is exactly their line of thinking: “Machines should be recognized as the inventors of their creations, but they shouldn’t own patents. Instead, the owners of the machine should get rights to the patent. Machines should not have patents, the researchers argue, because they ‘do not have legal personality or independent rights, and cannot own property.’”[10]
There is, however, a flaw in this logic. As an article by Rose Hughes in IPKat points out, the machine owners are happy to acknowledge that they do not deserve credit for the invention, but they are willing to accept the commercial benefits of ownership. As Hughes writes: “[w]hilst Dr Thaler insists that he is prohibited from listing himself as an inventor of the applications “because he has not contributed to the conception of the instant invention”, Dr Thaler maintains that he should have a right to the algorithm’s inventions. If another party were to use the algorithm to invent another invention, Dr Thaler and the team at the University of Surrey maintain that this invention would belong to Dr Thaler. But how does Dr Thaler derive this right from the algorithm inventor?”[11]
Evolution of patent law
In order to acknowledge the role of artificial intelligence in the creation of an invention, while also maintaining a level and fair competitive playing field in an economic sense, it suggests that some revisions or tweaks to the patenting system ought to be considered. For instance, it may be that inventions developed by AI bestow different rights to the owners of that invention — similar to the situation with Standard Essential Patents. Maybe there should be a right for others to license the benefits of the AI invention on fair and reasonable terms.
When devising such rights, policy makers should consider a fair return for the use of open, public data sources and the potential use of open source code in the development of the algorithm. Inventors should be prepared to describe in much fuller detail the mechanics of their ‘Creativity Machines,’ so that bodies can make fair judgements on where the economic benefits should lie and for whom.
The accelerating pace of innovation
Regardless of how the situation is resolved from a legal standpoint, AI promises a future of much faster innovation. This challenge is not new — R&D teams have long been seeking ways to optimise innovation processes so that new inventions can reach the market faster. The use of AI accelerates the urgency of dealing with this challenge. Patent offices can expect patent applications to arrive at their door at an ever-faster rate. AI will intensify activity in situations where it is used to replace tasks that were previously carried out manually, such as the auto-writing of patent applications embodied in services provided by organizations such as Specifio. Specifio describes the ability to transform two to three pages of claims into first-draft patent applications. As they explain, “These auto-drafts typically include 20–30 pages of text plus basic system and method figures. All that is left for the practitioner is adding in specific examples and explaining any unclaimed details or additional figures.”[12]
The growing body of patents also means there is more chance of contest and litigation. It will have a massive impact on the workloads for legal and patent specialists, inviting questions such as:
· Are the current processes going to be able to cope with the demand for this changing situation?
· Should AI be used to accelerate parts of the patent application, trademark application or other legal processes?
· If so, where should human intervention begin and end — what are the boundaries that should be set?
A challenge will be to define the relationship between humans and technology and to get that balance right.
Increasing efficiency
Whatever the answer to these questions, today’s immediate requirement is for patent application and legal processes to become much more efficient. Patent offices and legal practitioners must therefore identify aspects of their workflow that could be expedited through tactical use of artificial intelligence. WIPO provides an excellent use case, whereby the classification of patents into their respective technology areas can be carried out by AI. As it explains: “IPCCAT helps patent filers and examiners in IPOs to automatically categorize patent applications into technical units according to their International Patent Classification (IPC) class, subclass or main group.”[13]
Discovering the right answers to these questions will demand insight and input from all relevant stakeholders across the world. To ensure that the most beneficial amendments are made, much more education is needed on how the philosophical (and from that starting point, the legal) viewpoint on IP is formed and maintained, so that all parties can lobby for an outcome that truly represents their interests.
About this article:
This article was written by PatSnap Academy, which provides free, on-demand video courses covering cover a range of topics relating to IP and innovation, presented by renowned industry experts.
Sources:
[1] https://www.bbc.co.uk/news/technology-49191645
[2] https://www.wipo.int/wipo_magazine/en/2017/05/article_0003.html
[3] http://imagination-engines.com/iei_history.php
[4] http://artificialinventor.com/wp-content/uploads/2019/07/Fractal-Container-Application.pdf
[5] http://artificialinventor.com/wp-content/uploads/2019/07/Neural-Flame-Application.pdf
[6] https://www.crowdjustice.com/case/intellectual-property-for-ai-inventions/
[7] https://www.bbc.co.uk/news/technology-49191645
[8] https://www.federalregister.gov/documents/2019/08/27/2019-18443/request-for-comments-on-patenting-artificial-intelligence-inventions
[9] https://limegreenip.hoganlovells.com/article/104/patents-law-on-employees-inventions-germany
[10] https://gizmodo.com/should-artificial-intelligence-be-credited-as-an-invent-1836920203
[11] http://ipkitten.blogspot.com/2019/08/the-first-ai-inventor-ipkat-searches.html
[13] https://www.wipo.int/about-ip/en/artificial_intelligence/search.jsp
