26 April 2018

Britain has the chance to lead the world in AI. Will we take it?


The government announced a “sector deal” for the artificial intelligence industry today, which it says is worth £1 billion. About a third of that comes in the form of public funding and a third from new private sector investment, the rest being a mixture of expansions and in-kind contributions.

Matt Hancock, the congenitally connected culture secretary, says: “Artificial Intelligence is at the centre of our plans to make the UK the best place in the world to start and grow a digital business. We have a great track record and are home to some of the world’s biggest names in AI like Deepmind, Swiftkey and Babylon, but there is so much more we can do.”

Britain really does have a crack at world leadership in this stuff, but it won’t be easy and this latest announcement is only the start.

Let’s count our advantages. We have a vigorous tech startup culture, unleashed by reforms to taxes for both entrepreneurs and investors brought in by the coalition government.

Last year about 600,000 new firms were created. Tech startups attracted more funding than in any other European country and early-stage companies attracted about £3 billion in investment, almost double the total for 2016.

It’s possible, in other words, that there is the sort of explosion of commercial and academic experiments going on here that places like Silicon Valley and Boston have experienced in the past. The critical mass of researchers and entrepreneurs in London, with lots of rich lawyers, doctors, accountants and oligarchs to sell their wares to, sometimes makes even Californians envious.

Britain’s video-game industry, for example, may sound trivial, but it’s surprisingly good at finding new ways of doing things that can be applied elsewhere. We’re a more digitally connected country than Germany or France and we (mostly) speak the American language.

Case in point: the Newcastle based but international accountancy firm Sage is a pioneer in chat-bots to help small businessmen. Sage envisages giving us AI robots to check our tax returns in future and its foundation has established a “BotCamp” for school-leavers, partly designed to tackle the significant skills gap that still exists.

As part of the deal announced today, Sage plans to expand this as an AI pilot programme. (Declaration of interest: my family holds shares in Sage.) Its London rival Cleo has a virtual financial assistant that learns from users’ data how to advise you on your finances.

The National Health Service gives Britain another possible advantage in AI. Some of the most exciting opportunities for machine learning are in drug discovery, diagnosis and the analysis of scans to help radiologists target tumours. Having all patients treated by a monopoly provider has its drawbacks, but it does potentially provide a gold mine of data for machines to learn from, which is missing in the United States. Too much of the NHS’s data has yet to be digitised, however, so it’s not all good news.

One if the firms touted by today’s announcement, Your.MD claims to have built the world’s first AI personal health guide that provides you with medical advice from the NHS via a mobile phone.

That health-related AI got off to a bad start when an early deal between Google Deepmind and the Royal Free Hospital came under fire for its handling of personal data could also turn out to be an advantage. Reacting to that episode, Britain may now become something of a pioneer in the ethical regulation of data.

Just as the Warnock Report on embryo research in the 1980s led to a sensible settlement that prevented culture wars over reproductive research and treatment (including the “other AI”, artificial insemination), so an early settlement of the ethical issues around data could be a British opportunity in AI.

Can we find a middle road between heavy-handed European state regulation of data, and the domination of the American market by big tech companies, with their inevitably monopolistic tendencies?

It won’t be easy but may be possible. China, which will probably be the world’s biggest AI investor, is unlikely to be as sensitive to individual rights.

True, compared with Germany or Canada, Britain’s AI scene looks a bit of a mess, with the Alan Turing Institute, the Centre for Data Ethics and Innovation and other institutes vying to be the DARPA or MITI of this technology revolution.

But that could turn out to be an advantage, too. As Mao Zedong said (before doing the opposite): let a hundred flowers bloom. That’s how we pioneered the industrial revolution, after all.

There is undoubtedly a lot of hype around AI at the moment, driven by the new ability of machines to learn without being imbued with expertise, and it is possible that it will all fizzle out into a handful of examples of slightly smarter software, nothing more. Previous AI dawns have proved false. But that seems unlikely this time.

Also, even if the technology is revolutionary, it is the people who adopt it – and that may be all of us – rather than the people who invent or provide it, who will enjoy most of the benefits, as was true with other pervasive innovations such as electricity.

But there could still be great benefits from an early lead in the development of AI. After all, Europe has yet to produce a digital giant on the scale of Amazon, Google or Facebook, and Britain is well placed to break that duck.

Matt Ridley is a journalist and author. His most recent book is 'The Evolution of Everything: How Ideas Emerge' (Fourth Estate).