Curated data, analytics and artificial intelligence (AI) have been driving innovation across many sectors. The Covid-19 outbreak has triggered an acceleration in its use in the development of digital health, from modelling the outbreak and predicting patient demand, to drug development and diagnosis. Intellectual property law will need to adapt to protect investment in developments and enable innovation.
Protection of data as an asset
Data has been and remains a key driver for growth. It is the fundamental basis that empowers AI. With more and more data being gathered and AI allowing more sophisticated analysis of large data volumes, the options for data-based business models and commercialising data are increasing. Digital services are often provided in exchange for data rather than a service fee. It is therefore no surprise that, as the importance of data as the "new currency" increases, so do the associated legal issues.
Rights in data
Legal protection for "data currency" has not yet followed this development. The law offers little protection to data as an asset unless steps are taken to protect it as confidential information. In the European Union, the latest addition to the legal toolbox for data dates back to the mid-1990s when EU Directive 96/9/EC introduced a sui generis right for databases. This right protects the investment required to collate existing data in a database. It does not, however, protect the creation of data and the data itself.
At the end of the last century, when the investment protection for databases was introduced, the internet was only about to surface. The compilation of data was cumbersome and associated with significant cost. With automation improving and reducing the costs of data compilation, this is less relevant in today's world. Awarding protection for the investment in collating data no longer reflects the commercial realities. The current focus on digitalisation and AI may therefore revive the discussion on affording protection to data as such, which has been less prevalent recently.
Data powering AI and the patentability issue
From a development perspective, the current lack of protection of data is positive and has enabled data to be widely used to make inventions with the help of AI. The speed at which AI can analyse data enables research and development that drives increasing numbers of scientific breakthroughs. Machine learning, robotics, expert systems and symbolic learning generate output that a person of "normal skill in the art" would not have been able to make.
From a legal perspective, this challenges the long-established principles on which patents are granted. If AI makes inventions that a human skilled in the relevant field would not have made, the thresholds for awarding monopoly rights to incentivise research and development need to be rethought.
AI-assisted research in healthcare
AI-assisted research is of increasing importance in the healthcare sector to repurpose known drugs to treat other indications. Five to seven years can be saved when repurposing a known drug compared to the development of a blockbuster drug. The risk of failure is much lower, too. It is therefore no surprise that 25% of the pharmaceutical industry's revenue is generated with repurposed drugs. They are also expected to play a significant role in treating Covid-19. Such low barriers to market entry have triggered extensive AI-assisted research for new uses of known substances.
Data is mapped to find connections between drugs and diseases. Algorithms are implemented to identify unknown relationships and interactions. Repurposing companies use AI to monitor drug pipelines for drugs that have disappeared, drugs that were approved but are no longer manufactured and drugs that were abandoned in the process of their development. More often it is not the pharma that identifies a second use but a third-party expert in AI-assisted research.
Protecting AI inventions
This use of AI opens a pandora's box of legal issues that depend on the specifics of the technology used: Was the specific AI better than the usual AI? Does the AI merely implement a specific search task or is deep learning part of it? Today, human inventors behind the AI are still easy to identify. It is, however, merely a matter of time until the machine will by-pass its creators and develop its own process for more efficient research and development. This will lead to complex questions of attributing inventorship or even awarding patent protection at all.
Legislators, patent offices, and courts will need to strike a careful and innovation-friendly balance between recognising the use of AI in inventive processes and awarding protection for it while at the same time not raising the bar too high, in particular for inventions that are not or only minimally based on AI.
The increasing role of AI in scientific breakthroughs is challenging the principles on which patents are granted. - Julia Schönbohm IP Partner, Germany
The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.