With Brexit done, and the UK now free from certain regulatory encumberments, the British government ought to be devoting some time to thinking about artificial intelligence. AI is, after all, one of the most important technological advancements of the future, but it wonâ€™t happen by itself. The latest update to Tortoiseâ€™s Global AI Index found that the UK is in danger of falling further behind the AI frontrunners, the US and China, and being overtaken by countries such as Canada, Israel and Germany.Â
Thankfully, the government does appear to be doing some of this thinking. It recently tasked the independent AI Council to come up with recommendations for the UKâ€™s national AI strategy. Those recommendations were published at the very start of this year, in the form of a document called the AI Roadmap â€“ and many are very wise. A commitment to skill-building; making The Alan Turing Institute a truly national body; and increasing public sector investment all are important, necessary proposals.
But if we are to truly overcome the challenges that the UK faces in commercialising and scaling our AI research for long-term success, then we must go further.
Consider skills. AI needs to be taught in later-stage education â€“ in research fellowships and PhDs â€“ but also in schools. Making AI a part of the curriculum, and perhaps even committing to a ten-year programme of AI training, would be a significant step in ensuring that the public are familiar with the benefits and uses of the technology. And that means learning some pretty fundamental things. As weâ€™ve seen during the pandemic, many people arenâ€™t as familiar with the concepts of probability and risk as we might have previously thought, yet both of these concepts underpin AI.Â
As big as that policy would be, there are some bigger moves â€“ culturally speaking â€“ that the government ought to be making. Thereâ€™s a general sentiment in the world of AI that data ought to be available freely, as data is what fuels the technology. However, as admirable as this sentiment is, it is also a little naive. Why should the UK be giving away its rocket fuel to competitors, with no sensible commercial regard? The US or China would never think like this, and neither should we. Or to put it another way: if we gave NHS data to companies to develop new products, should we not make sure that, wherever those products are made, the NHS gets the best possible financial settlement for them?Â Â Â
And then thereâ€™s what is perhaps the governmentâ€™s best tool for encouraging the development of AI â€“ and something that the Roadmap underplays â€“ regulation. AI products are new and operate in the real world. Autonomous vehicles, personalised medicine, fintech. These are areas where adoption will be critically dependent on regulation, legal precedent and even seemingly mundane considerations such as insurance. This is an area where the government can actively create a fertile environment. There would not be airliners in the skies if former governments had not done this in the past.
Besides, the government can do more to help in the sphere of regulation than it can in the realm of funding. There is, it should be noted, no shortage of private money available to budding entrepreneurs with impressive technology and a good business idea: London-based startups raised more than $10.5 billion in funding last year, with a typical pre-IPO round reaching ÂŁ80 million. By contrast, the British Business Bank, the government-owned bank dedicated to developing smaller businesses, contributed ÂŁ200 million to help fund high-growth businesses â€“ a mere drop in the ocean. In truth, the government is no better placed to pick winners than venture capital firms are, and, as a steward of taxpayersâ€™ money, it is certainly under more scrutiny.
Public sector investment has, in some cases, actually been detrimental to technological development. Innovate UK, the public body designed to help businesses grow through innovation, is an ambitious and forward-thinking idea, but it has too often supported the progress of tech companies that fail to generate value or achieve commercial success, thereby perpetuating a culture of grant dependency rather than product innovation. Similarly, the massive EU Quaero project, intended to take on Google, not only wasted vast sums of money, but also created such a shadow that it basically hid the really innovative European players.
If the UK is to truly declare itself part of the AI arms race, greater consideration must be given to the commercialisation of AI research in general. Thanks to the efforts of previous governments, the UK now has a vibrant community of start-up and mid-size companies, but the final piece is missing. There are currently no deep technology businesses â€“ a category of business that includes AI startups â€“ in the FTSE100. The reality is that when our scale-ups are on the verge of having a real impact on our economy â€“ typically when they are valued around the $300-400 million mark â€“ they are snapped up by overseas capital and the prospect of a more tech-friendly NASDAQ listing, taking the benefits of their work elsewhere. We have to keep some of these companies in the UK, not by legislating it or through public funding, but by creating an environment where they can become $10 billion, $50 billion, or even $100 billion world-class companies.Â Â Â Â
This means sweating both the big stuff and the small stuff. Darktrace, the first company to apply AI to the challenge of cybersecurity â€“ and which was supported by my own company, Invoke Capital, and is due to list in London later this year â€“ was born out of a collaboration between British intelligence agencies and Cambridge University mathematicians. But the importance of the more prosaic things that ought to be considered when scaling a business for success, such as creating a customer support service or writing product manuals, cannot be underestimated.
The AI Roadmap was an important step in the UKâ€™s ambition to become a â€śscience superpowerâ€ť â€“ and one that I was pleased to see taken. But if we are to meet the governmentâ€™s aim of achieving a ten per cent increase in UK GDP through AI by 2030, we must nurture our science base, our institutions, our R&D, and our pipeline of ambitious AI companies. And crucially, once we have them, we have to create the right environment for them to stay here.
Dr Mike Lynch OBE FREng FRS is an entrepreneur and founder of Invoke Capital.