What to know about COVID tracing apps

I have a group chat with some high school friends, who were recently talking about the COVID Alert CT App, a new voluntary state-sponsored COVID contact tracing option. Here's how it works, according to the Hartford Courant:

The app uses Bluetooth to detect when another person with the same app comes within 6 feet. Your phone exchanges a secure random code with the other phone to record that they were near. If the app senses that you've been with 6 feet of someone for a total of 15 minutes or more in a day, the app adds their phone's random code to a list of close contacts. Devices that are just passing by or stay more than 6 feet away from you are not included on the list.

About 300,000 Nutmeggers signed up for the app on the first day it was available; two days later, nearly 20 percent of the population of the Constitution State had signed up.

This is, largely, good. But it may not be good enough. As Thorin Klosowski wrote at Wirecutter:

Setting aside the efficacy of the technology, adoption is also key to these apps' success. One study from Oxford—which also included authors from The Wellcome Trust Center for Ethics and Humanities, IBM, The Alan Turing Institute, and others—suggests that an adoption rate of 56%, with individuals over 70 years of age self-isolating, could help control the spread of the disease significantly, but lower adoption rates will still help.

[…] 

Once a person is notified of an exposure, they need to be able to get a test, to record in the app if they test positive, and to self-quarantine, all steps that can be difficult. The ACLU's Gillmor notes, "If they don't act on it, if there's no quarantine, then there's no point in the notification in the first place."

A low adoption rate could also lead to false negatives. Ashkan Soltani, an independent privacy and security researcher, notes that an app not detecting an exposure doesn't mean much. "It means no one that was tested positive had a phone that had this app installed," Soltani says. And he worries that it may give people a false sense of being virus-free.

(full disclosure: I also write for Wirecutter)

Klosowski also noted some of the quirks of the contact tracking through Bluetooth or GPS:

Bluetooth and GPS weren't made for this function, so quirks in the technologies may prohibit them from being effective. Bluetooth signals move through walls, for example—if you're in an apartment building and your neighbor tests positive, you may get an exposure notification even if you haven't seen your neighbor face-to-face. This possibility can make Bluetooth unreliable as an indicator of contact. Apps also can't know if you're wearing a mask or if you spoke to the other person through plexiglass. A paper published in Plos One challenged Bluetooth's effectiveness for contact tracing and showed that Bluetooth's distance measurements were so inaccurate on a tram—throwing both false positives and false negatives for contact—that the app wasn't any more useful than notifying people at random.

There are, of course, valid privacy concerns about contact tracing apps — which is particularly concerning when you consider that their effectiveness is still in question. Of course, the best way to improve that effectiveness is to get more people to sign up for such a service, which is difficult when there are valid concerns about effectiveness and privacy, even if the app itself comes from a place of a good intention.

Communities working together are the best way to overcome this pandemic. But you should be fully informed about what you're doing before you make your decision either way.

COVID Contact Tracing Apps Are Far From Perfect [Thorin Klosowski / Wirecutter]

University App Mandates Are The Wrong Call [Gennie Gebhart, Jacob Hoffman-Andrews, and Andrew Crocker / EFF]

Image: Public Domain via Pikist