Are AI Clusters a key to overcoming the UK productivity gap?

By Stephen Allott and Maurice Glucksman

The UK has a well-documented decades-long 20% labour productivity gap with the US, Germany and France.  With Brexit potentially creating extra challenges in overcoming this competitive disadvantage, the UK could fall even further behind.  But if you are an optimist it could also be a moment for the UK to take a giant leap forward as it has done in the past when confronted with no other choice.


Britain has been very good at doing this recently but it could be even better and Brexit could galvanise a new era of even more productive collaboration between public institutions and private business.   There are many sectors of the UK economy that have bright sparks of world class performance but McKinsey Global Institute Productivity research shows very clearly they are held back by vexing bottlenecks.  We believe this can be overcome. We decided to begin working to advance some already very strong work on creating industry clusters.  We could have chosen any one of many industry sectors to provide a foundation for our thinking but as it happens we already have a cluster of knowledge, connections and momentum in the AI space, we are taking a stab at working on how AI could be a backbone technology to drive more rapid development of many high performing industry clusters in the UK and indeed a high performing cluster of AI suppliers in the UK.


We believe the rise of Artificial Intelligence could be a springboard for rapid productivity growth in two ways:  AI is a functional industry unto itself and AI is also an increasingly fundamental element of any high performing industry from high tech to Entertainment and even Art.   A strong pure AI cluster is already developing in London spurred on by an imperative to develop tools for Fintech, Drug Discovery and many other applications.  The basic functionality of these tools is transferrable to other industries and the reverse is true also: AI applications designed for specific situations can be generalised to provide the backbone for applications in other industry sectors. We think this is a model the UK should support with a big coordinated push from the public and private sectors.  


However, it should not be done indiscriminately and it should be done in clusters because spread too thinly the net effect will be minimal and there has to be a well thought through drive to support strong clusters that can raise the game to world class levels and then float all other boats in its wake.  Why else are clusters important?  If you measure the productivity of two businesses running exactly the same processes but one is in close proximity to other related businesses and another is isolated, chances are the first one is going to produce more than the second one and scale up faster.  It is well documented that clusters co-located with leading research universities enable iterative development of ideas and an easy flow of people from academe into local industry. Furthermore, research by Sherry Coutu shows scaled up businesses are much more productive in spawning new startups.  Clusters create powerful virtuous feedback and AI can act as a turbocharger across many sectors.  

Its not clear to us if the priority should be to push hard on a pure AI focused cluster and then fan out the inventions to industry sectors or the other way around.  In our experience with the adoption of new technologies it has worked better the other way around: you solve a specific problem, create a tool, like a hammer, and then hunt with your hammer looking for nails.  But that’s not the way it worked with Nuclear Power or with the discovery of the Higgs Boson, and we think its just unclear.   So we think the best strategy is to support both and in fact that’s what’s happening anyway, but so far we haven’t seen it articulated explicitly nor have we seen policy being developed on this precise premise.


But we think it is widely accepted that setting out to develop clusters is not going to work just because you throw money at them.  The right conditions must be encouraged and Brexit has thrown up a number of red flags that could inhibit rapid development of an AI Cluster or AI driven industry clusters.  The most obvious is talent.


There is now a fierce global war for AI talent and skills and in fact a number of high profile AI companies in the UK, like Deepmind, Swiftkey and  Magic Pony and have been acquired precisely for this reason. The worthy winner of the Future of Britain competition for the EU Green Card idea may be very significant. If implemented there needs to be a selection criteria and decisions about who gets the Green Cards. 


But it cannot be assumed that the War for AI talent will be won by the UK just by opening the door.  We’re leaning heavily on McKinsey here but it happens that extensive research has been published on the The War for Technical Talent.  It’s a complex topic but to sum it up the war for talent is only won if talent are attracted as well as allowed in.   The UK has had many things going for it as a raw talent attractor, but again Brexit throws up another red flag: is top talent really welcome?  Will the UK invest in the best infrastructure and the best education and the best conditions for families? Are the best future opportunities and amenities in the UK?  We think the answer is obvious: amongst others, people with key high value skills and key amongst them: Artificial Intelligence (AI) researchers, entrepreneurs and practitioners must be attracted to and retained in the UK.  How to do that is a major issue.


But, lets make an assumption for the moment:  the talent issue is solved.  Let’s look at the opportunity for AI adoption in business. The UK has a number of world-class industries where AI could be a source of advantage such as Financial Services including especially Insurance, Trading and Asset Management.  And the UK has several world-class industries where AI is starting to penetrate, e.g. Drug Discovery, Advertising and Legal Services. The solutions to closing the 20% productivity gap are many faceted but we believe that when we complete our work we will able to say that with rapid adoption of AI, the UK can close the gap or even vault into the lead. What measures can Government take to accelerate adoption? A step change increase in AI courses at leading universities and investment in research can produce more trained AI practitioners. We could try a Korean style national plan. There could be national competitions like the one for the maritime longitude clock. A weekly or daily BBC, ITV, Sky or Amazon magazine programme on AI or national conference on AI could all help. Many measures should be tried.


Second, let’s look at the AI suppliers. The demand created for AI in the UK is already fostering the growth of UK based AI suppliers. With 10 out of the top global 100 AI companies, the UK is the AI leader in Europe but currently sits in a distant second place to the USA.  Arguably London is already the home for at least one AI supplier cluster with the critical mass that enables positive feedback loops operate and reinforce competitiveness. London needs to be marketed as a great place to found an AI company to attract internationally mobile entrepreneurs.  And there are other places in the UK where AI driven clusters are forming but may need a push to reach critical mass:  Cambridge for Biotech; Manchester for Digital Marketing and Sports Analytics; Birmingham for Business Services and Betting; Edinburgh for Fin-Tech; Glasgow for e-commerce; Sunderland for Factory Automation and Robotics; Bristol and Bath for Education and Virtual Reality; Leeds for Gaming and Simulation and Belfast and Reading for Cybersecurity and this is just scratching the surface (apologies to all those clusters forming and not acknowledged here)


All 3 elements: Talent, AI adoption and AI suppliers must come together as a cluster with critical mass to really drive the industry forward. And a fourth plank is pure research.  It’s no accident that a vibrant cluster around Biotech has formed around Cambridge University and most other emerging clusters have a strong research Institution in close proximity.  How policy makers can catalyse the growth of a cluster is obscure and many policy initiatives focus on building networking organisations. What you actually need is successful businesses and perhaps some way to measure and managed the performance of clusters.  We’ve even debated the radical idea of appointing a Chief Cluster Officer for each cluster in the UK who would be an advocate, much like a mayor is an advocate for a city.  Having large successful companies whose staff found spin-offs helps. Having direct competitors in close proximity also helps. Knowledgeable investors and flexible office space can help as well.


So far Brexit has been a disruptive force but the search for ways to make a virtue of Brexit has created a window of opportunity to really shake things up.  Britain could be a world leader in AI research and applications and we think aggressively supporting AI industry Clusters could be a key to achieving world productivity leadership. 


There is window of opportunity open now and we believe the UK must dive through it before it closes.

Published in

Leave a reply

Thank you! Your subscription has been confirmed. You'll hear from us soon.

Log in with your credentials


Forgot your details?

Create Account