Who Will Win the European AI Race?

What was initially a rumble has become a cascade of opinion, research, and policy; advanced Artificial Intelligence (“AI”) is coming, and it’s going to propel the Fourth Industrial Revolution. Each major EU economy has announced its desire to become a global AI leader. On one hand, this refers to the development of a thriving AI sector; on the other, it corresponds to a country’s ability to diffuse innovations into its wider economic infrastructure, enhancing prosperity through substantial productivity gains. We are interested in how Britain’s AI ecosystem and future strategy compares to that of its competitors, in particular, France. This is a broad topic as a country’s AI advantage can be analysed through a multitude of characteristics - this blog will focus specifically on data infrastructure.

Regarding the development of AI, big data underpins success. Machine learning enables systems to learn without programming, but this application relies on data. The more data it is given, the more accurate it becomes. Thus, the availability of data is central to AI-driven economic growth. Furthermore, it is national data infrastructure - the amalgamation of policy, processes, and institutions - that facilitates this availability. So, how do France and the UK compare?

According to a study by the Centre for Data Innovation, the UK significantly outperforms France across a range of availability indicators. The public sector is highly data-intensive and, rather interestingly, France has more limited availability due to a lack of data sets. This in itself does not explicitly disadvantage French AI researchers and developers. The open data of all EU member states is centralised in the European Data Portal. However, in contrast to France, the progressive and streamlined nature of the UK’s open data policy forms the core of an ecosystem that fosters AI innovation.

A further key challenge to AI development relates to barriers to data sharing in the private sector. Data published by varied organisations enables businesses to innovate in diverse industries. However, private sector data is often personal or commercially sensitive and, as such, cannot be made open. As part of the AI Sector Deal, the British Government, academics, and business leaders have agreed to an agenda that explores mechanisms such as Data Trusts for data-sharing between organisations. A French government report has also identified the pooling of company data as an essential component of data policy.

So how do the two states compare? Both countries have identified private sector data sharing as a priority area, and both states are in the early stages of formulating a decisive strategy. However, as is illustrated by the Centre for Data Innovation, the value of the UK’s data economy far exceeds that of France. Where France is ranked 18th, the UK is ranked 3rd. The data economy refers to the size of the market for trade in data. This is important as it reflects the extent to which institutions are already in place to enable companies across industries to share data and create new value through innovation. The UK’s performance on the measure of data economy illustrates that not only are the characteristics of its data ecosystem superior to that of France, the cultural conditions to support data-sharing between organisations already exist.

Where the UK is at a disadvantage relates to the outcome of the Brexit negotiations, hindered access to EU data would undoubtedly reduce UK competitiveness. The UK will leave the Digital Single Market, but the Government has emphasised its desire to have a strong data relationship. Presumably, this means a high degree of regulatory alignment, which in theory suggests a seamless flow of data. However, a key issue relates to data localisation. While the EU intends to liberalise the flow of data within its borders, it could introduce EU-level data localisation requirements, which would require third-country digital businesses to relocate part of their operations to the EU.

To conclude, the UK outperforms the other leading EU economies in data infrastructure and, as such, is well placed to win the European AI race. However, the key question going forward is the extent to which leaving the EU Digital Single Market will inhibit this.