Covid-19 – Contact tracking: proper preparation is the key
We see a lot in the news these days about a mobile application that can be used to detect carriers of Covid-19. Many people are feeling rather uneasy about the concept, however. We look at some of the potential pitfalls and possible solutions.
The greatest concern the technology community has with digital contact tracing or tracking applications is how effective they are. The data collected cannot be considered significant unless the majority of the population uses the app–a study from Oxford University says about 60%. Since data are only collected when certain people come into contact – those with cell phones who are using the application – we know nothing about the others. Under such circumstances, a false sense of security could be created in the population, and that would be dangerous.
Also, we need to know who is transmitting the virus and who is not to make the application truly effective. Carriers who have no symptoms are not tested. The app can’t really work unless there is a massive screening program.
Finally, transmission of the virus does not only take place from person to person. It may also be transmitted via infected surfaces. And it’s impossible to build that factor into the development of the application.
The planned technology is said to protect privacy to a high degree, as the data are anonymized and not stored in a central location. In fact, the information gathered is not correlated with geographic data, since the identifier changes frequently. And it’s only stored on the phone.
However, this kind of operation, while safe, is not perfect. For example, information that makes it possible to identify a person who has the virus and track his or her contacts will be saved on a server, which means anonymity is lost. Besides, it’s hard enough to maintain anonymity in everyday life. If an individual only comes into contact with one other person in the course of the day and that person turns out to be a carrier, the first person will know who it is when told he or she has come into contact with an infected person.
Social pressure down the line
It’s crucial to fully determine the end use of the application: do we want it to tell us who should be put in quarantine, or simply help screen for the virus? There’s a big difference between the two, with different consequences for the population.
Contact tracking applications could also play a role in marginalizing certain groups. Would people who don’t use the app but have been carriers or live in a hot zone risk being denied access to stores? If the answer is yes, that would mean marginalizing low-income seniors and others living on a low income who don’t have cell phones, as well as many essential workers and members of cultural communities. Are we really trying to make their lives even harder? We need to think about the collateral effects of the application.
No turning back
I’m also concerned that contact tracking applications will eventually be considered trivial and routine. If the pandemic continues for a long time, as seems likely, people will have time to get used to it. And after the pandemic, it’s possible that this technology will be used in other contexts, for example in police investigations, to find out who was in contact with whom. This would not be the first time that a technology that was originally extremely useful wound up being used under questionable circumstances – just think of the use of facial recognition in the United States. Once new surveillance measures are put in place, it’s very hard to put the genie back into the bottle.
Avoiding the worst-case scenario
Right now, I don’t believe it’s a good idea to develop contract tracking apps, because there are too many slippery slopes. Before this type of application is put in place, it’s crucial to have an open debate. One point that is frequently raised when an application is created is that apps should be created not for groups, but with groups. That’s the only way to make sure the app really does what we want it to do.
The people who are currently thinking about these issues are doubtless very smart, but they may have a few blind spots. We need to make sure that groups with relevant and diverse expertise have an opportunity to contribute. This can be done quickly, by simply creating efficient structures and asking the right questions.
It’s important for government agencies and companies working together on the project to be transparent. The app is not a solution in and of itself – it’s part of a toolbox. We need to know what else is in the toolbox, and whether the app is demonstrably relevant in that context. Also, accountability mechanisms need to be fleshed out. And most important, all this needs to be considered as part of the app design process and in governance practices that are put in place.
Clear parameters need to be established and be made known to everyone right from the word go. We need to know where data can be found, how they are protected and how they are destroyed. Finally, since the app won’t have been tested before we start to use it, structures and mechanisms must be planned to ensure that the app does not create social disparities or inequalities.
This is an extract of an interview conducted by Maryse Guénette.