Artificial intelligence is making an impact in real-world situations, from agritech to music composition, and healthcare data analysis to customer service. As the technology expands, this is an important moment for establishing a competitive advantage.
Leading AI business Google DeepMind is home to an exciting array of projects. Meanwhile, it is building an impressive patent portfolio. International patent applications number 41 at the time of writing (search result here). These applications form a strong starting point for an international portfolio – international applications offer an efficient route to patenting in multiple countries as compared to filing separately in each one – and a few have already progressed to granted status in the US and China.
A global race
In the global race for AI leadership, the US, China and Europe are each vying for the top spot. Currently, these regions show strength in different areas. The Artificial Intelligence Index 2018 measures data on a range of activity including publications, entrepreneurship, University enrolment, and patenting. The US leads on patent filings, but Europe is consistently the largest publisher of papers – a 28% share in 2017. Numbers of people accessing robotics operating system software are about equal in Europe and the US, with China not far behind. Robot installations annually are particularly strong in China, with numbers growing rapidly.
As this growth accelerates, voices expressing concern about where AI will take us in the future also multiply. Back in 2014, renowned physicist Stephen Hawking warned about the threat that would be presented by AI that outstrips human capability.
More recently, tech billionaire Bill Gates compared AI to nuclear weapons, offering both peril and promise to human society.
With these concerns in mind, Europe is taking an ethics-led approach. The EU's High-Level Expert Group on Artificial Intelligence brings together academics and industry representatives to help address these issues, and develop law and guidelines for ethical use of AI. The group published draft ethics guidelines for consultation in December 2018 with the final version expected soon.
The Guidelines are expected to be a living document, acting as a voluntary code for organisations developing and using AI. They build on two components to trustworthy AI:
- respect for fundamental rights, applicable regulation and core principles to ensure an ethical purpose
- technical robustness and reliability.
The Guidelines address specific areas of concern that AI raises:
- identification without consent – automatic face recognition for example
- covert AI systems, where a human does not know if they are interacting with a human or a machine
- mass citizen scoring
- lethal autonomous weapon systems.
It recognises that new dangers are likely to emerge as the technology evolves, and the guidance will need to track and reflect these developments.
The group is also tasked to produce policy recommendations in May 2019. We will report on these when they appear.
Will principles hold back development?
Some criticise this ethics-centred approach, arguing that it will hold European industry back. But it seems likely that greater recognition of individuals' rights will not be confined to Europe. EU laws and guidance may exert pressure for similar measures elsewhere. Following the introduction of Europe's data privacy reforms (the GDPR) in 2018 new privacy laws have been developed in other countries (India and Brazil, for example). A new Californian Privacy law will take effect in early 2020.
It's worth noting that the G7 meeting in Canada in December 2018 produced a Common Vision document, with all members signing up to principles for ethical development. It is interesting to see that this is something that US legislators are also taking seriously, with plans for new guidance under development.