Beware Dominic Cummings’s technology investment arms race

The Cummings view of the ‘AI arms race’ and the type of policies needed to compete in it is strikingly similar to the sort of stuff one heard about the technological prowess of the Soviet Union, writes Ben Chu

Dominic Cummings has particular view on AI (REUTERS)

If the off-the-record briefings are to be believed – something not to be taken for granted with this Government – Downing Street believes a no-deal Brexit will be a price worth paying for the sovereign ability to grant copious amounts of state aid to new British technology companies.

The irony of a Conservative government adopting the economic Brexit rationale of hardline left-wingers – that Britain must be free to intervene in the economy at will - has been widely noted.

And the answer to the winner pickers in Downing Street is the same as was offered to the Lexiteers: look a little closer at the European rules and also what actually happens on the Continent.

Just as the single market level playing field regulations are manifestly no bar to the national ownership of utilities and rail operators and the like they are also compatible with subsidies and support for UK technology firms.

European regulations make it clear that state aid for research and development projects is permissible provided it’s for “fundamental research”, “industrial research” and “experimental development”. That’s rather a large loophole for any government minded to exploit it.

It may be true – as those who have worked in government attest - that UK ministers and civil servants and legal advisors have often cited the supposed constraints of European state aid rules as a convenient way of batting away proposals for subsidies and support that they simply don’t like but that’s a reflection of the preferences of UK policymakers rather than the constraints imposed by European rules.

To read a blog post of Boris Johnson’s chief advisor Dominic Cummings from last year – and presumed briefer-in-chief now - on these matters is to read not forensic analysis of why EU state aid regulations stifle EU technological advancement but a wild rant about how “Brussels spends its time on posturing, publishing documents about ‘AI and trust’, whining, spreading fake news about fake news, trying to damage Silicon Valley companies.”

Cummings complains about the fact that the UK government allowed Google to purchase the London-based artificial intelligence company Deep Mind in 2014. But, again, this decision was nothing to do with the EU.

And we should note that Germany’s economy minister last year unveiled plans to allow the Berlin government to intervene to prevent a non-EU company acquiring a stake of 10 per cent or more in any German firm involved artificial intelligence (AI), robotics, semiconductors and biotech.

Germany, it would appear, does not feel that such an interventionist approach is incompatible with EU membership.

If Cummings’ blog is the extent of the case for rejecting any EU state aid rules it’s no wonder that the Government isn’t publishing its own vision of the UK’s post-Brexit state aid regime and technology sector interventions. We would probably see a distinction without a difference. Either that or we would see a difference that would rightfully alarm.

Because what’s also notable about Cumming’s blog is also the admiration he displays of the policies of the autocrats of Beijing when it comes to this area, approvingly citing an article which claims  “China’s national AI strategy poses a credible challenge to US technology leadership.”

But this isn’t the only perspective. There is no question China is channelling considerable state resources into its AI sector along with many other forms of industrial research and development. But some close observers, such as Oxford’s Eric Ding, think the Chinese AI advances are overhyped, as one would expect from a regime steeped in the methods of Communist propaganda.  

The fruits of China’s expensive efforts to build a domestic semiconductor chip and electric vehicle manufacturing sector have also been disappointing.

The Cummings view of the “AI arms race” and the type of policies needed to compete in it is strikingly similar to the sort of stuff one heard about the technological prowess of the Soviet Union during the Cold War.

“There is a large body of Soviet scientists, engineers, and production men, plus many highly developed subsidiary industries, all successfully directed and coordinated, and bountifully financed,” wrote the influential American journalist Walter Lippmann in 1957, after the Soviets successfully launched their Sputnik satellite in 1957.  

“The United States and the Western World may be falling behind in the progress of science and technology.”

We now know that Soviet science was, even then, fatally wounded by a political system, which valued ideology over truth and which responded to dissent with persecution.

No one knows what the future of technological development will be in China’s repressive and anti-privacy political system.

But history suggests authoritarian regimes do not tend to produce high-quality science nor technological breakthroughs - certainly not breakthroughs for widespread consumer use in complex market economies.

The Leninist undertones of Cumming’s approach – the veneration of science and mathematics combined with his authoritarian personal political style – have been remarked upon. But if he’s willing to reject a free trade deal with UK’s largest trading partners for the sake of a Cold War-style technology arms race the unflattering comparison may go still deeper.