What are we talking about when we talk about contact-tracing apps?

The Australian government has introduced a contact-tracing app called COVIDSafe. In this guest post Dr. James Parker, a Senior Lecturer at the University of Melbourne's Law School asks what is at stake.

...........

COVIDSafe, Australia’s new ‘contact-tracing’ app, was released at 6pm last night, a matter of hours after two state governments announced they would be reducing restrictions on movement and gatherings. It’s hard to imagine the timing was a coincidence. Indeed, the government had already signalled that the app would be one of three key pillars to Australia’s ‘exit strategy’. By morning, the media was reporting it had already been downloaded more than a million times nationally, news that would surely put a smile on the Prime Minister’s face, the Guardian said. 

Whether these initial downloads translate into ‘users’ and for how long, is another matter, however. Indeed, with so many outstanding questions around the app’s functionality, we still don’t know what ‘using’ COVIDSafe really involves, or how this will affect the reliability of the data it collects and extent of the roll out.

As expected, the conversation is already pitting public health against privacy.  One friend tells me via WhatsApp that Google and Facebook already reap far more of our data anyway, so ‘if it helps, what the heck, I’ll download it’. Someone on twitter says they’d much rather take the advice of ‘doctors’ over ‘privacy activists’. Whatever happened to ‘collectivism’ protests another. According to Greg Hunt, the Health Minister, downloading the app is a lot like buying ‘war bonds’, only cheaper. ‘What we’re doing in fighting this fight is we’ll be asking people to download an app which helps us trace the virus quickly and the more people who do that, the more we can get back to a more liveable set of arrangements.’

There are already a number of major assumptions here, the most important of which is that the app will actually work if enough people download it. There has been far too little questioning of this basic premise, given that – as far I know – there is no empirical evidence to suggest that it will, and when the one thing we do know is that it failed in the only previous test-case (Singapore). 

What would ‘working’ even mean in this context? Keeping infection rates low? Keeping them low under the current emergency measures? Or also as restrictions start to lift? Even if numbers do stay low (and how low is low anyway?) and the app is used to contact and test some of those at risk, will we consider the app a success only if those people couldn’t have been found via tried and tested interview-based contact tracing? Or if digital contact tracing ends up doing the same thing, only cheaper? Is the app cheaper? To what extent is COVIDSafe about public health and to what extent about economics?

It strikes me that there’s a basic falsifiability problem here. There is simply no way of knowing whether this app has been effective. Whereas what’s certain is that it will help to further normalise state surveillance and/as ‘technological solutionism’, and that any perceived failures will be blamed, in the first instance, on uptake levels and ‘privacy activists’ undermining civic values in a time of crisis.

There are other problems: 

1)    The argument ‘from big tech’ - ie that our data is being mined for profit anyway, so we might as well donate it for the benefit of public health – is precisely the wrong way around. The unrelenting expansion of ‘surveillance capitalism’ (Zuboff), platform capitalism (Srnicek), data colonialism (Couldry and Mejias) – pick your synonym – needs to be resisted. If it instead becomes the benchmark against which all privacy questions are measured, that battle is lost.

2)    The idea that doctors, epidemiologists, or even scientists in general, are the ones we should be listening to in the conversation about COVIDSafe fundamentally misunderstands the nature and scope of scientific knowledge. This is just another version of the point made by Jeremy Baskin recently about the rhetorical and political power of being able to say that the response to Covid19 is being ‘guided by the science’. Science may be able to help us understand the benefits of contact tracing. It may even be able to found an argument for some kind of app. But it can’t tell us whether the data collected by that app should be centralized (as in the case of COVIDSafe) or dispersed, how long it should be held for, what legal consequences should attach to an alert that you’ve come into contact with an infected person, what to make of the app’s enabling legislation (which will presumably be debated in Parliament at some point… though God knows when), or – more fundamentally – whether the government can be trusted to use the data it collects for purposes other than the ones it tells us about… Because as Bernard Keane argues in Crikey, the question isn’t so much whether COVIDSafe is a threat to your privacy, but whether the government is. And you don’t need to be Barnaby Joyce or Pauline Hanson to have concerns on that front.

Perhaps, unlike Joyce and Hanson, you’ve got nothing to hide. But as Lizzie O’Shea, the chair of Digital Rights Watch, points out, you would hope that the government didn’t either. And if it doesn’t, why not release the app’s source code? As indeed the department of health’s own privacy impact assessment recommended. Or, for that matter, release the infrastructure at the government’s end: the ‘whole data custody chain’? Some of the source code, two weeks after the app’s release, which is what we’re now apparently being promised, is simply not good enough.

These are no longer conversations about an app. And they have very little to do with public health, or even really privacy. They are about political accountability and transparency; about fantasies of technological fixes to social problems; and more specifically about how digital technology – and tracking and surveillance technologies especially – are increasingly imagined as essential to society’s proper functioning, despite the scant evidence to this effect… not to mention the political consequences.