From Boston to Silicon Valley, technology companies are building innovative tools to track the spread of the COVID-19 virus by monitoring the smartphones of infected people. But these “contact tracing” systems may do little to control the disease, due to circumstances that even the biggest corporations and smartest engineers can’t control.
The biggest problem is there has been so little testing that there are not enough people to trace with smartphone monitoring.
As of Wednesday morning, fewer than 1 percent of US citizens had been tested, according to the COVID Tracking Project. That means there are probably many thousands of infected people who are invisible to any tracking technology.
“None of this works without ubiquitous and frequent and highly distributed testing," said Omer Tene, vice president of the International Association of Privacy Professionals in Portsmouth, N.H.
Moreover, any tracking method that maintains a high level of privacy protection relies on people using it voluntarily, and it’s likely that some infected people would not participate.
Also, people who don’t have smartphones don’t have an obvious way to participate. And because a Bluetooth-based system can detect infected people who are too far away to pose a risk, the system could generate lots of false-positive warnings, leading some people to stop using it.
Still, the recent alliance between rivals Apple and Google on COVID-19 tracking technology shows the tech industry believes there is enough potential in such a system to mount an enormous effort to bringing it online.
The Apple-Google system would rely on the short-range Bluetooth radios built into every smartphone. They can be directed to transmit a digital “key” to other nearby phones. These keys change at regular intervals, do not provide any information that could identify the phone’s owners, and do not reveal the locations of the phones. The key-swapping shows only that the two phones — and presumably their owners — were within about 20 feet of each other, the range of a Bluetooth radio.
People who have tested positive for COVID-19 would be able to transmit a record of their Bluetooth interactions to a remote server, including the keys from every phone the infected people have encountered. The server would then generate a message to the owners of those other phones, warning them they had recently been near an infected person, and suggesting that they get tested.
The Apple-Google system, which is supposed to be launched by mid-May, is similar to the Private Kit system under development at the Massachusetts Institute of Technology. With Private Kit, a user installs a phone app that tracks his or her location throughout the day. People infected with the coronavirus can then choose to share their location data, which is displayed on a map so healthy people using Private Kit can see if they were nearby. Private Kit does not reveal the identities of infected people.
Even with privacy protections, these app-based systems are fraught with problems, such as persuading enough people to use them. For example, in the island nation of Singapore, about 1 million people have downloaded a government-sponsored tracking app. That’s less than 20 percent of the population of 5.6 million. A Singapore government official told the Straits Times newspaper that for the app to be fully effective, it must be used by about three-quarters of all citizens. Persuading hundreds of millions of people in the United States to participate would be a massive challenge.
Apple and Google are planning to include their tracing software in their iOS and Android operating systems later this year, eliminating the need for a separate app. But consumers may choose not to use the service, because it might generate a flood of false-positives, warning of infection hazards where none exist.
Tene noted that a Bluetooth radio signal can carry for dozens of feet and penetrate walls, and the Apple-Google system won’t know exactly where the two people on either side of an encounter were situated — whether, for example, they passed each other in separate cars, or were in adjacent hotel rooms.
This kind of loose proximity, Tene said, could mean “you will get pings for being in proximity to a sick person without really being at risk.”
Think of two strangers walking past each other at a distance of 20 feet. If one is infected, the other might receive a smartphone warning, even though there’s almost no risk of infection at that distance. Now imagine this happening a dozen times a day, as more people test positive for COVID-19.
“If people start being inundated with red flags and warnings," Tene said, "it will water down trust in the system.”
And that could lead millions of phone users to simply ignore the warnings.