“It’s not a failure, but it didn’t work. Just four months after its launch, the application for “tracing” positive coronavirus cases, StopCovid, was officially buried by Emmanuel Macron during his television interview on October 14. Downloaded only by 2.6 million people in France, StopCovid will only have sent 472 notifications to potential contact cases, according to government figures released last week.
But the executive intends to correct the situation with a new application. Entitled “All Anti-Covid”, it should see the light of day on October 22. What lessons can be learned from this failure and why has StopCovid not achieved the desired results? Olivier Blazy IT lecturer and responsible cybersecurity master at the University of Limoges, answers the questions of 20 Minutes.
Why has the StopCovid application failed in its primary mission, namely to alert and prevent the increase in contamination?
The application did not produce the expected results because too many hopes were placed in it in relation to the means deployed. It was to be the “Holy Grail” and limit the number of contaminations when it is only a tool. And this tool was to complement many efforts, as the CNIL recalled when it was launched. It was to come as a reinforcement of the hospital response and the rise of screening tests. But we found ourselves in cases where the rare people alerted by StopCovid sometimes had to wait several weeks to be able to be tested.
Then, the government’s communication focused on security and the protection of personal data came up against alerts from the scientific community. And beyond the low number of downloads, the technology was chosen to track our interactions – Bluetooth – is not necessarily suitable *. Several recent studies have highlighted the fact that the detection of the distance between users remains very approximate.
Is it a French peculiarity?
Compared to some European neighbors, France really has a catastrophic record from the point of view of application downloads. But beyond that, even massively installed tracking apps have shown their limits. This is the case in Iceland, for example. While nearly 40% of the population installed the app, authorities recently said that the impact on the spread of the epidemic was not enough to continue rolling it out.
In five months, more than 2 million people have been identified and called by the Social Security “brigades” responsible for “contact tracing” , including 550,000 “patients diagnosed positive” and 1.5 million “contact cases”. Is a digital tool really essential in view of these rather positive human results?
I would tend to say no. To develop such a tool, you have to take the time to reflect. Illness is something intimate, and putting technology in there can be a drag for people. When a person learns bad news, in this case, that they have coronavirus, I don’t know if their first instinct is to enter a code in the application to declare themselves positive.
Obviously, it must remain a voluntary act, and it is very good to have opted for this solution, but the fact is that it is rarely a reflex. Perhaps it would have been necessary to imagine a “booster shot” by telephone, for example, to encourage patients to report on the application once the diagnosis has been made. There, the fact that it only happens through technology, it does not help “digest” the announcement of the disease. Whatever the machine does, it takes people behind it, especially to call the contact cases.
What would be the necessary conditions, in your opinion, for the future “Tous Anti-Covid” application to work, where StopCovid failed?
More information needs to be given to users. People are worried about the epidemic. Tell them, through this tool, that their region is targeted by specific devices or that they can go to such and such a structure to be screened if they are in contact, that may be good. When users activated StopCovid, they had no indication of the real-time operation of the application. They did not know how many people had been detected thanks to the application for example, but that can be a reassuring indicator.
One could also imagine that the tool gives the user the number of his daily close contacts. Depending on this figure, we could adapt our behavior. In my opinion, we have not sufficiently worked on the “reassuring” role that the application could have in this context. These developments simply require a different approach to “user experience”. Obviously, there is no question of saturating the application with unnecessary buttons. When we already see that the Prime Minister was not even able to download the application, it is not worth developing a tool too complicated.