The UK recently handed Google, a foreign company from a country the UK no longer entirely trusts, the keys to its national data treasury because it is afraid of hackers.
China, meanwhile, is moving with characteristic speed to centralise all the data it can lay its hands on, and then lock that away from foreign eyes, to conjure up a competitive advantage in artificial intelligence (AI).
The reality is a little more complicated than that on both sides, of course, but for the purposes of strategic decisions facing the likes of South Africa, you can boil down the UK and China's approaches to their places on a two-axis graph: risk aversion vs opportunity seeking on the X, and cross-border cooperation vs data sovereignty on the Y.
For countries that have not yet decided between those options, it's a little like China and the UK is putting on an educational pantomime in which they act out the more extreme options available for the edification of their peers.
And you'd better believe many countries are watching closely, with no consensus yet on what will be the winning combination.
The UK headlined its zero-cost partnership with Google as an effort to get rid of legacy systems and get its civil servants AI-fit. But you didn't have to listen too closely to catch the fear underlying the deal. The technology minister, Peter Kyle, spoke of the "vulnerable, archaic servers" where public institutions keep their data, "vulnerable to outages and to cyberattack".
The UK is rather defence-minded right now, as it tries to spin up a plausible defence to possible future foreign aggression, but it is also extremely aware of current vulnerabilities thanks to the three-month saga that has been the M&S hack. That chain of stores closely resembles Woolworths in many respects, including the general air of management and technical competence it carries – and its estimate for the cost of a ransomware attack that crippled some of its services is currently north of R7 billion.
With a history of criminal records being stolen from the Legal Aid Agency this year, attacks on the National Health Service, and the British Library forced to use something like half its reserve funds to recover from a 2023 attack, the UK government is just waiting for the big hack to hit at the heart of its operations.
Its response? Google Cloud. For now in a vague, non-committal way that needs not go out on tender, but it is clear where things are going. Fellow US data giant Palantir got its teeth into the British health system through just such a no-(initial)-cost deal, which rolled over seamlessly into a mega contract.
Handing all your data to Google has an echo of the old "nobody gets fired for buying IBM" to it. When we discover that a nation state – almost certainly the US – has a backdoor into Google's cloud, the impact will be so widespread that politicians won't be whipped for the decision to trust it. That's perhaps the best part of doing enterprise business with a household name that consumers trust with everything from their e-mail to their photos: we're all in this together.
There is, however, both political and practical risk in handing data over to a company that can be pressured by a rogue US administration, no matter how loudly it proclaims it can offer geofencing for sensitive data.
Still, at the stroke of a pen and without spending a cent (so far), the UK has plugged into one of the best security teams in the business, in a way that gives it access to data-hungry AI tools that should be not too far behind the cutting edge. While, let's face it, almost certainly contributing its data to the training of shared models.
But it will be contributing to a shared good, to tools that will be available, at a price, to everyone not actively sanctioned by the US.
China, by contrast, is playing its cards close. With DeepSeek, it proved it can fast-follow US AI developments without the need for collaboration. Its government is increasingly treating data like a natural resource of strategic importance – something can put it ahead in any number of fields as long as it can out-produce everyone else when it comes to data. China rightly has faith in its production capacity, and it has a battle-tested approach to export restrictions via rare earth minerals. It is not hard to chart the strategic course from here.
The obvious problem with China's in-progress sovereign data ocean is that oceans are hard to secure against both poachers and poisoners. To the extent it can keep the rest of the world out, though, China will both hold back models elsewhere and have the chance to create something unique.
South Africa can hitch its wagon to China, or the US, or it can see what Europe eventually gets up to, or it can go for the pan-African approach again. It is not yet under significant pressure to choose what to do with its AI-friendly data, and hence has not been under pressure to decide on a strategy.
That is an enviable position to be in, but it is also a vulnerability. It could result in SA making tactical decisions based on options put in front of it, rather than based on principle. What do you do when Google puts a zero-cost deal in front of you, or when China gets gruff about a Digital Belt and Road initiative, or when the US takes a with-us-or-against-us approach to data, as it is with de-dollarisation?
Right now we have no sense of where SA is likely to go. But at least it gets to watch the immediate consequences for and reaction to the UK and China's opposite directions.
Share