Comment

Ministers must back AI specialists now – or lose us

A woman looking at a wall of deas
Britain doesn’t have enormous tech companies. But we do have the talent and the entrepreneurialism Credit: Hero Images

About a third of the leading machine-learning and AI specialists who leave the UK’s top institutions are now working in US tech firms, according to research carried out by The Telegraph. But how concerned should we be and what, if anything, can we do about it?

In some ways, the US tech giants are doing us all a favour. The academics they hire are usually allowed to continue their research (with access to the vast resources of a tech company) and, because top talent is in such demand, the researchers are able to insist on their work being published.

But, while this is of benefit to science, it undoubtedly means that a portion of the economic benefits of their work is lost internationally.

The solution, then, is not to keep these people in academia but to offer them the opportunity to do their science in a way that would benefit the UK economy – inside a UK tech giant.

But the truth is that, at least in the post-2000 era, the UK doesn’t have enormous tech companies. We have the talent, which is why the big US tech firms come shopping here, and we have the entrepreneurialism.

But time and time again, the best UK companies end up being acquired by larger companies from overseas, whether its DeepMind bought by Google in 2014, or Magic Pony bought by Twitter in 2016. So how can we ensure there is a British Google?

We can learn a lot from what happened in China. In the Nineties and early 2000s, everyone looked to the West Coast of the US to understand hyperscale start-ups. People claimed that there were many factors: engineering expertise, the unique Silicon Valley ecosystem, deep pools of capital, strong universities and a risk-taking culture.

Recently, though, some of the world’s largest companies have come out of China: Didi Chuxing and JD.com, for example. Many of the factors claimed to fuel the US – often tied up in a sense of Silicon Valley exceptionalism – were absent. What the Chinese companies did have was an enormous local market to expand into and a lot of government help.

A US or Chinese company can grow from zero to several tens of millions of turnover rapidly, at which point the considerable costs of starting an international operation become viable.

We do things very differently in the UK. But I believe that, like the Chinese, we can and should see government procurement as a strategic asset.

With government spending representing about a third of UK GDP, the Government could turn the technology sector into a world-beating force by strategically procuring UK SMEs. This will be hard at a time when UK public services are stretched by spending cuts, and Brexit is taking focus.

But investing in AI presents a fantastic opportunity to make public services better for less – to improve productivity while nurturing a strategic national industry. In an era when only 2pc of burglaries are solved, we need a step change and AI’s ability to spot patterns in data has a big part to play.

Google CEO Sundar Pichai speaks during a product event in San Francisco
Silicon Valley giants, such as Google, have been investing heavily in artificial intelligence Credit: Eric Risberg/AP

There has been an aversion among civil servants to working with 
 start-ups for fear they will fail. However, Singapore and Norway have shown it is possible for governments to act more like venture capitalists, using sovereign wealth funds to back innovative new players. The US too does this within the AI, data and national security space through In-Q-Tel. Why not the UK?

The Government has made some moves. The headline figure of £725m invested in the Industrial Strategy Challenge Fund programmes sounds like a lot but it is spread thin across all areas of research and development.

There is also the £20m GovTech Fund to encourage tech businesses to come up with innovative solutions for more efficient public services. But this is tinkering around the edges when you consider annual revenues for Google alone totalled $110bn (£85bn) last year.

The Government should be establishing more funds outside of departmental budgets that mid-stage AI companies can bid for. This would bring more innovative and creative solutions to public-policy problems.

Historically, UK government has been unable to favour UK companies because of EU procurement laws. It is probably no accident that the EU single market has failed to generate a Google or Facebook equivalent. Brexit offers a chance to start using our government procurement in our national interest.

At ASI Data Science we have one of the biggest teams of data scientists and machine-learning researchers in the country, of whom 80pc hold PhDs from top universities; exactly the kind of people Google and Facebook are interested in.

We get acquisition inquiries from almost all of the large companies that we talk to. So far we’ve resisted. But eventually, someone will make an offer that I owe it to our team to think very seriously about it.

The Government has a very real chance to keep my company, and companies like us, independent. But the timescale for action is short. Within the next few years, the currently molten AI landscape will harden into winners and losers, and we will have lost the opportunity to influence the outcomes.

If we don’t act soon, the organisations that have the scale and cash will make offers that it would be irresponsible to refuse.

Then another chance will have been blown, another independent British company will fold into a US giant, and we’ll have to wait at least another generation for the possibility of a British hyperscale company.

Marc Warner is the chief executive of ASI Data Science     

License this content