After months of work, the UK has abandoned the way its coronavirus tracking app works, causing a guilt game between the government and two of the world’s largest tech companies. So what went wrong?
In late March, I received a message from a leading figure in the UK technology sector. This person said they were helping the NHS “on a very substantial project that will be launched in a few days and potentially save hundreds of thousands of British lives.”
It was the first I knew of the plan to build a contact tracking app, a project that soon seemed to be at the center of the government’s strategy to defeat the coronavirus and help us emerge from the blockade.
The tech luminary somehow assumed I could be a project consultant – I made it clear that it couldn’t be my role, but I was very interested in following his progress.
Now, almost three months later, after missing the deadline after the deadline, there has been a radical change in direction. The app that has been developed so far has been demolished and a new approach based on a system created by Apple and Google will be tried.
But there is no guarantee when, if anything, this will be implemented. So what went wrong?
When the NHSX digital division team was reunited, he was told that they were engaged in a vital mission. According to a presentation the team was shown that the Covid-19 app would have four objectives:
- Stop or slow the epidemic
- Check the flow of patients in hospitals
- Help people return to normal life
- Collect secondary data for use by the NHS and strategic leaders
Once installed on a user’s phone, the app used Bluetooth to keep track of other people with whom they had come into close contact, as long as they also installed the app. So when someone tested positive for the virus, alerts were sent to their close contacts from the past few days telling them to go to quarantine.
Epidemiological experience was provided by a team of Oxford scientists who claimed that there was an urgent need to identify the people who were spreading the virus without knowing it. “Very fast contact traceability was probably essential,” says one of the Oxford team members, dr. David Bonsall. “And smartphones have the technological ability to accelerate this process.”
But using Bluetooth on smartphones to track contacts was an untested technology. However, the team was inspired by Singapore, which had released its Trace Together app using that system.
But soon it became clear that using Bluetooth was complicated. Reports from Singapore suggest that people were reluctant to download the app because it always had to be open on the phone, draining the battery.
So on April 10, a surprising announcement came from Google and Apple. The two tech giants – on whose software virtually all smartphones in the world depend – said they would develop a system that would help Bluetooth contact tracking apps work smoothly. But there was a problem: only privacy-focused apps could use the platform.
Apple and Google preferred decentralized apps, in which correspondence between infected people and their contact list took place between their phones. The alternative was that the correspondence was done on a central computer, owned by a health authority, which would end up storing a lot of very sensitive information.
The app that the NHS was developing was based on a centralized model, which Oxford scientists believed was essential if the health service was able to adequately monitor virus outbreaks.
Two days later, with a huge fanfare, health secretary Matt Hancock unveiled plans for the Covid-19 app, promising “all data will be handled according to the highest ethical and safety standards and would only be used for treatment. and the search for the NHS “.
But immediately privacy activists, politicians and tech experts raised concerns. “I recognize the overwhelming force of public health arguments for a centralized system, but I also have 25 years of experience where the NHS is incompetent in system development and repeatedly violates their privacy promises,” said Professor Ross Anderson of Cambridge University.
However, the project was still catching on with the first trial version of the app at RAF Leeming, Yorkshire. The trial took place in artificial conditions, with military men and women placing adjacent phones on the tables to see what happened.
Meanwhile, privacy-conscious Germany has become the last country to switch its app to the decentralized model, using the Apple and Google systems. It appeared that Apple had made it clear that it would not cooperate with a centralized app.
Michael Veale, a British academic who works with a consortium that develops decentralized apps, warned that the NHS app was on the wrong track, asking on Twitter “the UK will push ahead with an app that won’t work on the iPhone – which has devastated adoption in Singapore? “
But the UK went ahead with an Isle of Wight trial. Just started, mr. Hancock told the public that they had a “duty” to download the app when it was available and that it would be crucial to have “our freedom” back when the blockade was alleviated.
The first view of the app showed that it was very simple, asking users if they had a fever or a continuous cough. But any symptom alerts sent to the contacts simply echoed the standard “be careful” advice: at this stage it was not possible to enter the test results in the app. It has left many residents confused.
However, the fact that the app was quickly downloaded by over half of the island’s smartphone users has seen the government successfully branding the process.
Meanwhile, the Financial Times has revealed that the government had hired a Swiss software developer to create a second app, using Apple and Google technology. NHS insiders quickly diminished the significance of this move, although one admitted that “Downing Street is getting nervous.”
Work continued on a second, more sophisticated version of the original app, which would have been tested again on the Isle of Wight before being launched nationwide, although the original mid-May deadline had been lost.
On May 20, however, it became clear that the government’s attention was shifting to the track of manual contacts. The prime minister announced that a “beat the world” tracking system would be put in place in early June, although issue 10 pointed out that the app’s contribution to the system would come a little later.
As May drew to a close the head of the larger test and track program, Baroness Dido Harding, said the app would be the “icing on the cake” of the project. It was no longer the cake itself.
At the beginning of June, other deadlines for the national release of the app had come and went. Three weeks after the trial began, the trial residents were getting restless, with very little information about how an updated version of the app was going or when it was coming.
France launched its centralized Stop-Covid app, which had sparked strong criticism from privacy activists and digital minister Cedric O said that 600,000 downloads in the early hours were “a good start”.
On June 4, Business Minister Nadhim Zadhawi was induced to declare that the app should have been ready by the end of the month, but that was the last fixed deadline that would have been promised.
Singapore, which has continued to struggle to make its contact tracking app work, has announced plans to offer all citizens a wearable device in the hope that it will do a better job than a smartphone.
On June 14, Germany became the largest country to launch a decentralized app on the Apple / Google platform. It quickly overtook France in terms of downloads with something approaching 10% of the population by installing it.
By now the silence of the British government on the NHS app was deafening. What was happening?
By lunchtime on June 18, everything became clear. The BBC broke the story that the government was abandoning the centralized app and moving on to something based on Google and Apple technology. Despite all the spin, the Isle of Wight trial revealed a disastrous flaw in the app: it failed to detect 96% of contacts with Apple’s iPhones.
The guilt game has already begun. Hancock and some scientists who collaborate with the NHS believe that Apple should have been more collaborative. Technology experts and privacy activists say they warned months ago how this story would end.
Apple claims it does not know that the UK was working on a “hybrid” version of the NHS coronavirus contact monitoring app using technology developed with Google.
In the meantime, there is little evidence worldwide that smartphone apps that use Bluetooth are an effective method of tracking contacts. In March, it seemed that the extremely powerful devices that most of us carried with us could help us get out of this health crisis. It now appears that a human being at the end of a phone is a much better option.