John Naughton

Repeat after me: there is no magic bullet for getting us through this pandemic. And smartphone-based proximity-sensing is definitely not that bullet, though it might be useful if two conditions are met. One is that it’s perceived by citizens to be trustworthy and protects their privacy; the other is that it’s deployed in conjunction with a massive increase in state capacity for testing and treatment. Neither condition will be easy to satisfy.

There are clear indications that the UK government is now actively considering use of the technology as a way of easing the lockdown. If this signals an outbreak in Whitehall of tech “solutionism” – the belief that for every problem there is a technological answer – then we should be concerned. Tech solutions often do as much harm as good, for example, by increasing social exclusion, lacking accountability and failing to make real inroads into the problem they are supposedly addressing.

The technology involved, though complex, in essence provides a way of automating what has been a way of dealing with plagues since at least the 1600s: find those infected, lock them away or treat them and then trace everyone with whom they’ve been in contact and quarantine them too. This is a very labour-intensive task that is not feasible in a society such as the UK’s. But many smartphones have low-energy Bluetooth sensors that automatically register the proximity of other similarly equipped phones, while most smartphones also log their location using GPS signals. So in principle we could use smartphones to do contact-tracing on a large scale.

That’s the principle. In practice, there are various ways of using these capabilities in the Covid-19 context. Centralised models involve phones equipped with an app to relay their data, supposedly anonymised, to a central server run by a government health authority. This may make things simple for the government, but it’s a nightmare in terms of state surveillance especially if the authorities try to make installation of the app compulsory.

Decentralised models involve keeping most of the data on your phone and only broadcasting to all the phones to which you’ve been close via a secure relay server if you’ve been diagnosed. All of your contacts’ phones will then inform their owners that they’ve been in contact with a diagnosed case of Covid-19. And of course all of the communications implied by this are encrypted by default. Because the individuals involved are notified immediately as soon as someone in their proximity is diagnosed, this method shortens exposure risk and enables health providers to suppress the virus rapidly. It restores agency to the individual, lessens the risks of state surveillance and better protects users’ privacy.

You don’t need to be a rocket scientist, let alone an IT expert, to realise that there are legions of devils in the details. (Harvard’s Safra Center for Ethics has a very good guide to some of them.) Who tells your phone that you’ve been diagnosed, for example? Given the possibility that – in a post-lockdown scenario – individuals with Covid-19 might be subjected to stigma, harassment or dismissal, they might be understandably reluctant to broadcast the fact.

Then there’s the problem that not everyone has a smartphone, even though it’s commonly supposed in tech circles that they do. The pandemic has revealed that a significant minority of the population (mostly older people) still relies on olde-worlde feature phones. Moreover, it turns out that not all smartphones are created equal: one estimate is that 50% of all smartphones can’t use the proximity-sensing systems being developed by Apple and Google. Given that any proximity-sensing system would probably have to cover at least 60% of the population to be truly effective, does this mean that Matt Hancock is going to be giving out Huawei handsets like Smarties to the Nokia-using poor?

I could go on but you get the point. The problem with magic bullets is that they sometimes miss their target. The biggest issue of all with smartphone contact-tracing, though, is that it would mark a step-change in state surveillance capabilities. Such a momentous decision cannot be left to Matt Hancock and his colleagues in their Downing Street bunker. This is a central point in a landmark review of the issue conducted by UK research group the Ada Lovelace Institute. A decision to deploy mandatory proximity-sensing technology, says the institute, is too important to be left to technocrats. There has to be proper parliamentary scrutiny and primary legislation with real sunset clauses. No fudging with orders in council by frightened ministers. I agree. If we get this wrong, not only will we not succeed in easing the lockdown, but we might also be kissing goodbye to the shrivelled democracy we still possess. There’s no lockdown exit through the App Store.

The Guardian, UK