Blog; Nuffield Council on Bioethics: Pete Mills, Assistant Director
“In order to make rights and laws function according to pure theory, the jurists place themselves in imagination in the state of nature; in order to see perfect disciplines functioning, rulers dreamt of the state of plague.”
– Foucault, Michel (1975) Discipline and Punish, pp.198-99.
For a week or so now the public discourse on the UK’s COVID-19 epidemic has been shifting from the horror of the daily death count to the hope that there may be a way of relaxing the restrictions that have been placed on citizens’ freedom of movement and assembly. The restrictions were exceptional and indiscriminate. Relaxing them will mean taking a more discriminating approach. What is at stake is enabling people to move about, interact and express themselves in different ways, to make friends, love and money. To enable this safely, a system of contract tracing has been proposed. Put simply, rather than locking down the whole population, contact tracing allows the selective isolation of confirmed cases and those thought to have be at risk of infecting others until their status is determined.
Contact tracing begins with individuals reporting symptoms suggestive of COVID-19, which may then be confirmed by testing. This triggers a process of identifying and contacting those they might have infected through prior contact. The people who are contacted can then be advised to isolate themselves and perhaps to take a test. If those people show symptoms or test positive the process can be iterated to contact any people they might have infected. Contact tracing has been used with success to arrest the spread of other infectious conditions such as HIV but, instead of ‘have sex with’ or ‘share needles with’, the criterion in the present case is ‘spend 15 minutes within 2 metres of’. Given that the latter can occur accidentally, say, on public transport or in a queue at the supermarket, the event and the identity of the other party may be harder to recall. Of course, as with HIV, it is possible to take precautions and avoid high-risk behaviours but, given the ease of transmission of the SARS-CoV-2 virus, ‘high-risk’ behaviours include many of the things that, hitherto, used to characterise everyday life and effective precautions may involve a panoply of personal protective equipment.
Both the scale and the speed of diffusion of SARS-CoV-2 infection make an automated system for contact tracing a superficially attractive tool. Such systems have been developed to operate via a smartphone app, such as the one under development by NHSX, rather than relying on human memory or laborious diary-keeping, intrusive interrogation of strangers and call centres full of tracers. (Surely, human tracers are as anachronistic as the human ‘computers’ who worked on ballistics research in the 1940s? Well, perhaps not.) But rather than as a possible adjunct to manual contact tracing, the app is now being advanced as a solution and, furthermore, one that is necessary owing to the occult nature of early infection and the speed of diffusion. (According to the British Society of Immunology and Academy of Medical Sciences expert group, it appears that asymptomatic spread accounts for 44-60% of infections, and peak infectiousness occurs 0.6 days before the appearance of symptoms.) This claim for the necessary introduction of the app was made, for example, in evidence before the House of Commons Science and Technology Committee this week (28 April) by one of the authors of a recent paper on the ethics of instantaneous contact tracing using mobile phone apps.
Necessity can mean different things. Clearly, in this case, it can’t mean that the contrary is impossible, that the app must be found in all possible worlds. It is, at best, a qualified necessity: the only means to a contingent end. The ends in this case are complex, however, involving trades-off between health and economic welfare, civil liberties and a range of other goods. We can argue about the ends. This, after all, is the very matter of politics. But, likewise, we can argue about the means. Granted, reaching an agreed end by alternative means may involve an intolerable cost. So perhaps necessary means: ‘necessary given the cost’. But different ‘costs’ are valued differently by different people. They cannot and should not be reduced to a single measure, such as economic value. This, too, is something we can debate.
There is a further sense in which something can be said to be necessary: if it is thought to be inevitable. There is a discourse, particularly within the ‘tech’ sector, that this way of managing an epidemic represents ‘progress’ and that progress is ultimately irresistible. This may be an assumption underlying ‘app solutionism’ but the structural conditions may already be more deeply embedded in infrastructures, systems and cultures. This is not always already inevitable but becomes so when we realise too late what is happening. The most likely reason for this happening is lack of transparency and debate. And lack of transparency and engagement has characterised the development of the NHSX app, which is now nearing deployment, while speculation has already moved on to possible ‘updates’ that will extend its functionality (for example, to upload potentially identifying location data or even to include facial recognition to validate immunity certification) and the desirability of retaining data for research purposes.
The app might well benefit subpopulations where people must come into sustained contact with strangers, for example, where people need to commute for work using public transport (it’s hard to think of a better capsule to incubate viruses than tubes, trains and buses, which is perhaps why London was so savagely hit) and leisure activities (theatre, concerts, sports events, which are also, of course, other people’s livelihoods). However, the app solution has well recognised limitations: it could be misleading (Bluetooth is a poor ranging technology with a wide margin for error), confusing (where multiple people use the same phone), gamed (people not carrying their phones in public or carrying others’) or maliciously exploited (where someone has control over another’s device or where they might use their device to cause another device user to self-isolate).
To contribute significantly to the suppression of viral transmission and the restoration of freedoms an app would need to be taken up and used by a large proportion of the population. This has been estimated at roughly 80% of smartphone users, or 60% of the public overall. This is one of the factors that led the Ada Lovelace Institute, in a recent (20 April) rapid evidence review, to conclude that “There is currently insufficient evidence to support the use of digital contact tracing as an effective technology to support the pandemic response.” It is, nevertheless, conceivable that the uptake of an NHSX app could be high enough to offer benefit as part of a strategy including manual contact tracing and testing. But because the possibility of anyone getting benefit from it relies on the fact that others have voluntarily installed the app, it requires resolving a problem of collective action. Herein lies the real problem.
The Government and its agencies seem to be shaping up to browbeat people into adopting the app solution. The NHSX Chief Executive recently admitted that it will be ‘tough’ to get the levels of uptake needed to make the approach work. He has said that ‘the message should be’ that the lockdown restrictions can be relaxed safely if the public embrace the app. He said that ‘an enormous comms effort’ will be needed from ‘all the people that people trust’ to get the app into all communities of the UK. We may already have seen this comms strategy piloted at the Downing Street briefing on 24 April, when Grant Shapps said: ‘It will be an NHS App. And I know how people feel about supporting our NHS!’ thereby associating frontline staff caring for people with COVID-19 and the people who brought us care.data. It is not surprising that there is already ambivalence. The Government’s approach betokens either institutional obstinacy, refusing to learn the lessons of the past, or a belief that ambivalence can be overcome by raising the stakes. It is not inconceivable that it will now lead either to the abandonment of the project or to somehow trying to make the app compulsory.
In Emerging Biotechnologies we cautioned against, on one hand, hypothecating societal challenges to particular technologies and, on the other, hypothecating technologies to particular challenges. The first obscures what alternative approaches might be available. (The Cambridge Centre for the Study of Existential Risk has recently – April 30 – published a solution scan of societal options to reduce SARS-CoV-2 transmission and spread that identifies 297 measures, albeit that many contribute marginally in combination with others). The second obscures the full potentialities and possible consequences of committing to a particular technological approach. (The latter could include, for example, ‘updates’ to repurposing of the app to incorporate immunity certification, extending use or retention of data to serve collateral purposes, and other forms of function creep.) We need to recognise how means and ends may be mutually implicated.
In the resonant phrase of Langdon Winner, ‘artefacts have politics’. There are a lot of issues, not limited to those of privacy, data security and information governance, but connected with social injustice and various forms of discrimination, that I have not addressed in this post, though they have been widely discussed by others and surfaced prominently at a roundtable meeting we held last Friday. While assurances have been given that equivalent provision will be made for those who either cannot access or decline to use the app, the implications of differentiating these groups do not seem to have been fully considered. This debate is the one that we should be having alongside the discussion of an ‘exit strategy’ (‘Covexit’?) from the current ‘lockdown’. (‘Exit’ is, of course, now a politically loaded term co-opting the support of ‘freedom-loving British people’ to whom the Prime Minister has often referred as his idealised version of the British citizen.)
The contact tracing app is a disciplinary technology. It enables micro-level surveillance and control in which individuals are ‘caught up in a power situation of which they themselves are the bearers’. It exists within a system of new disciplines to which citizens are subject that draws its justification from the value of protecting public health. In a democratic society, the reason to submit to such a discipline is for the good of others, not because a government prefers to arrange and deploy its citizens in a particular way. The app story so far is not so much a failure of design but a failure of process.