IP strategist: Will artificial intelligence replace IP strategy?
Artificial intelligence is already affecting the way we conduct R&D and may even change the way we think about invention. But it will only make the job of the IP strategist more important
Just because DeepMind’s AlphaGo artificial intelligence (AI) beat Go master Lee Sedol in 2016 does not mean we should be counting down the days until humans are replaced by machines. However, the technology behind AlphaGo will undoubtedly have a big impact on all kinds of industrial domain, with intellectual property being no exception. In April the Licensing Executives Society International (LESI) organised a panel on Artificial Intelligence: New Frontiers and New Challenges, which speculated on ways in which AI might affect inventions and intellectual property. These musings provide a good starting point for thinking about how this new technology could change the everyday work of IP strategists.
Artificial intelligence and invention
Stephen Boyers of IBM Watson has first-hand experience of how AI can help. Watson is already assisting with drug development and discovery, and the same cognitive solutions have been deployed to uncover new approaches to patient safety. Boyers has no problem contemplating that Watson will invent a drug within five years.
Julian Nolan of Iprova pointed to a quote from Eric Schmidt (former CEO of Google): “Between the birth of the world and 2003, there were five exabytes of information created. We now create five exabytes every two days.”
Similarly, while there were 1,000 scientific publications a year in 1900, this figure rose to 500,000 by 2000. In a world where patenting is a race against time and the inventor with the priority date often takes the spoils, AI is useful for combining various sources of information in novel ways – although there will always be a role for humans to review that information.
Artificial intelligence and the law
Eran Kahana of Stanford University accepts that the law’s attitude to AI will not stop the technology advancing, although it may create barriers to adoption. Using autonomous driving as an example, there is no shortage of predictions that by 2025, cars and trucks will not require human drivers. However, this overlooks the crucial fact that as yet, there is no legal framework for attributing liability for accidents which occur once a vehicle is driving itself.
At the LESI discussion, Francesco Zacco of the European Patent Office (EPO) reminded the audience that patent law requires a human inventor and that there is no likelihood that law and practice will change within the next five years. However, he also pointed out that machines are already compiling data for submission in support of a patent application and the EPO is considering how AI can help with the identification of prior art.
Artificial intelligence and monetisation
Mike Mulica of Actility (and former CEO of Unwired Planet) was struck by the abundant inefficiency associated with licensing patents. Unwired Planet was built on foundational technologies developed by OpenWave which enabled mobile devices to connect to the Internet. This capability turned out to be extremely popular. It is a recognised source of disappointment that litigation is currently the only accepted basis for price discovery. In Unwired Planet v Huawei (April 4 2017), Justice Birss offered a potential ray of hope, where better data would enable parties to understand the universe of relevant patents and the proportional share owned by each party.
Artificial intelligence and IP strategy
These views provide a snapshot of the capacity of AI to influence a variety of different IP-related endeavours. But what impact will it have on IP strategy itself? There are insights to be gained from the Aistemos IP Strategy Report. In this, Tony Clayton (former chief economist at the UK Intellectual Property Office and CEO of the Open Register of Patent Ownership) commented: “If there was better quality information about who owns intellectual property, that would provide a welcome boost to licensing and corresponding reduction in litigation.”
This neatly sums up where we are today. IP strategy is a company’s ability to integrate IP issues into mainstream value creation and risk mitigation. AI is the ability to harness machine learning algorithms and vast (and inexpensive) computing power to enhance existing activities and to unlock fresh insight.
If AI improves the treatment of illness or accelerates the rate of invention, we should all be delighted; though there will always be those who are convinced that AI threatens job security. A recent PwC report, for example, predicted that 30% of UK jobs could be affected by automation by 2030. A better view (endorsed by the same report) is that this will be offset by the creation of new jobs.
There are those today who direct their expertise to ensuring that a patent is granted or to successfully litigating a patent claim. Imagine a world where AI delivers IP business intelligence to enable companies to understand whether it is right to apply for a particular patent (or whether the patent application will proceed to grant), or to help an inadvertent infringer to avoid the dispute in the first place.
We are fast approaching this point. More IP data becomes available all the time and many of us are harnessing the potential of AI to deliver better information to the multi-disciplinary teams which are responsible for managing R&D, innovation and IP strategy. It is somewhat ironic that the world of intellectual property, with its roots in creativity and invention, has been so slow to adapt to accommodate this potential.
AI will not replace IP strategy; it will enable it to become mainstream.