Boston University proposal would involve tracking randomly generated IDs via short-range broadcasts as means to contain coronavirus infections
Researchers from Boston University have proposed a means of using mobile phone location data to alert individuals who may have been exposed to the novel coronavirus, without compromising those individuals’ privacy.
Such a mechanism could be used to underpin a targeted quarantine regime, which could reduce economic and social disruption as the coronavirus continues to spread around the world, said researchers Ari Trachtenberg, Mayank Varia and Ran Cannetti.
At present, quarantine measures are governments’ only option for containing the virus – and preventing healthcare systems from being overwhelmed – until a general immunity has built up in the population or a vaccine has been developed, they argue in a new paper.
Targeted quarantine measures can mean less disruption than broad lockdowns, but in the countries where they have been introduced, such as South Korea, Taiwan and Singapore, it has come “at high cost to the privacy of their citizens“.
Those countries’ systems involved centrally cross-referencing the precise mobile phone location histories of all citizens against the identities of infected individuals, in order to track down those who may have been infected.
“This potentially deeply personal data has then been released with only ad hoc partial ‘anonymisation’ measures that have left the private information of both infected and non-infected individuals vulnerable,” the researchers wrote.
The outcry against such an encroachment on individual privacy at a national scale risks the rejection of such tools in Western countries, they said.
And yet, such measures could be invaluable for diminishing the disruption caused by the coronavirus – particularly in its second or subsequent waves, the researchers argue.
“After quarantines are lifted…, we will once again see exponential infection rates… unless infected people can be very aggressively isolated, together with any potential contacts,” wrote Trachtenberg in a blog post.
As an alternative, the researchers proposed a technique that could use short-range broadcasts, such as Bluetooth or near-field communications (NFC) and would not involve disclosing users’ identities.
A proposed app would broadcast a random identifying number via one of these technologies, which would be sensed by the same app running on nearby smartphones.
The identifying number would be changed at regular intervals, such as every 1 minute or 1 day, and the app would keep track of the ID numbers it uses and those it encounters.
When a user tests positive for Covid-19, they could voluntarily share their list of randomly generated IDs with an agency such as the Centres for Disease Control and Prevention (CDC), which would make the numbers publicly available.
The app would regularly check the CDC’s list of numbers, and if it found a match, it would indicate the user may have been in contact with a person infected with the coronavirus, indicating the potential need for a test.
The approach would involve sharing only random numbers, with no obvious link to personal information, and would be entirely voluntary, Trachtenberg said.
Difficulties would include publicising the app and getting enough people to use it.
A study by Oxford University last week indicatd that at least 60 percent of the population would have to sign up to a given virus-tracking scheme for it to be effective.
“We believe that the privacy guarantees provided by the scheme will encourage quick and broad voluntary adoption,” the researchers said in the study.
“When combined with suﬃcient testing capacity and existing best practices from healthcare professionals, we hope that this may signiﬁcantly reduce the infection rate.”
At present, an opt-in virus-tracking app is reportedly being developed in the UK that could launch within weeks, while the US has a variety of different projects across multiple states that are making efforts to collaborate.
The EU’s Pan-European Privacy Preserving Proximity Tracing (PEPP-PT) initiative is bringing together 130 researchers across eight countries around a Bluetooth-based tracking scheme.
All existing initiatives have limitations, however – with one being the question of how to bring in the involvement of individuals without smartphones.