Illustration of a headset hovering over a map of the United States with coronavirus hotspots highlighteds

Health agencies across the U.S. are recruiting tens of thousands of people to make phone calls.

The hiring spree is part of an urgent push for contact tracing, the largely manual task of reaching people who may have had a brush with a coronavirus carrier so that they can isolate themselves and get tested. California is training 20,000 people to do this, and New York City plans to hire 2,500 in the coming weeks. Contact tracing is critical to disrupting the spread of the pandemic, and safely relaxing business and school closures, according to public health experts.

But Silicon Valley has promised a different future. Since the pandemic began, tech companies have proposed proximity-tracking apps that would automatically tell people if they’d come near someone who later reported having COVID-19.

More on Coronavirus and Technology

Apple and Google set out to lead the effort, and Wednesday the companies jointly released software tools public health agencies can use to build the apps. Since the program was announced in April, the companies have said their system would be easy to use, operate on almost every smartphone, and protect everyone’s privacy. Once they roll out, the apps could cut back substantially on the size of the army of healthcare workers needed to work the phones.

That might not happen anytime soon, however, according to interviews with current and former public health officials. Proximity tracking apps have a chance at beating back the pandemic only if they are coupled with widespread testing and are used by a large percentage of the public—about 60 percent, according to recent modeling from Oxford University. (If fewer people use the apps, they may still help prevent new infections, but the smaller the number, the smaller the effect.) That kind of widespread uptake is unlikely in the U.S. in the near future, experts say.

Proximity tracking apps make “implausible assumptions” about adoption, says Tom Frieden, a former Centers for Disease Control and Prevention director who is now the CEO of Resolve to Save Lives, an arm of the global health nonprofit Vital Strategies.

“Right now, they are a distraction,” he says.

Several public health agencies in the U.S.—plus 22 abroad—are working on apps that will use the system that was rolled out today. In a series of statements accompanying the announcement, health experts lauded the potential of the Apple-Google system to speed up contact tracing.

“It’s still early days, but app-based exposure notification can potentially play an important role in local or national test-trace-isolate intervention strategies,” said Trevor Bedford, a scientist at Seattle’s Fred Hutchinson Cancer Research Center.

Apple and Google say they agree that broad adoption is critical, and are taking steps to help. They are restricting the use of their technology to a single public health agency in each region and are implementing strong privacy protections to boost consumer trust. Additionally, they say the apps are meant to augment the work of human contact tracers, not eliminate the need for them. 

But these new apps face an uphill battle. Even before today, developers in some states had come up with homegrown alternatives, unhappy with some of the restrictions the tech giants have imposed on their technology—and so far, the adoption rates have been discouraging.

In North and South Dakota, a system originally developed to help college football fans find each other on the road has been repurposed for coronavirus contact tracing, with the support of both state governments. It’s being used by 4.3 percent of North Dakotans and 2.1 percent of South Dakotans, according to Tim Brookins, the app’s developer. Utah’s public health agency hired a social media firm, Twenty, to cook up a location tracking app for the state. The company says its app has been downloaded by about 1.5 percent of the state’s population since it was released last month.

These apps are fairly new, and you could argue they haven’t had time to gain momentum. But the picture overseas is sobering, too. Adoption rates range from 25 percent in Singapore and 20 percent in Norway to well under 10 percent of the population in India, where the national contact-tracing app has been made mandatory. Iceland—a homogenous country of 364,000, where fully one-third of the population lives in a single city—has one the highest rates at about 40 percent. 

In the U.S., about 30 percent of people say they would be willing to use an app for proximity tracing, according to new Massachusetts Institute of Technology research led by Baobao Zhang, a political scientist. The real adoption rate would probably be lower, she says. 

Adoption aside, there are two more reasons not to count on proximity-tracking apps to play a major role in the near term, according to public health experts: questions about their accuracy, and the critical importance of other factors—especially widespread testing—that are still missing.

Two Tracking Technologies

Contact tracing apps generally draw on two smartphone technologies: GPS and Bluetooth. But they both have weaknesses that could limit their effectiveness for this purpose.

The Apple-Google system relies solely on Bluetooth, the same short-distance signals that let your phone talk to your car’s infotainment system or to wireless headphones. Apps that use this system keep an anonymous log of the other phones you came within range of, and hold on to the data for 14 days, in case the owner of one of those phones reports that they tested positive for COVID-19. The apps won’t access GPS to record where you go, to help protect your privacy.

The apps in Utah and the Dakotas use Bluetooth to record which phones you are near—but they also keep a GPS log of where you go, which you can share with health officials if you test positive for COVID-19. The idea is that if you find out you’re sick, a human contact tracer can more easily track down other people who crossed paths with you over the previous two weeks.

Experts say both technologies have accuracy issues. Two people using Bluetooth apps that are sitting on opposite sides of a wall or a window, for example, could mistakenly register a potentially dangerous encounter because Bluetooth signals travel through barriers that the coronavirus cannot. (Apple and Google say that their technology can be tweaked to ignore weaker signals and cut down on false positives, but we won’t know how effective it is until the apps have been available to the public for some time.) And GPS can be spotty in cities and indoors. 

Online reviews of Care19, the app being used in North and South Dakota, suggest that many users have run into accuracy problems. Several recent reviewers said the app, which is currently rated 2.9 out of 5 stars in the iOS App Store, marks incorrect locations and misses some movements. Some say the app stopped working altogether. “I’m trying to help in contact tracing and this is doing nothing,” one wrote on May 4.

Brookins, Care19’s developer, says the app has seen major improvement over recent weeks in response to users’ feedback. But, he says, people should expect to edit the app’s results because it will inevitably get some locations wrong.

Privacy advocates have warned against using GPS to track physical locations for contact tracing, preferring less invasive Bluetooth proximity tracking.

But a single-minded focus on technology would ignore some fundamental issues that won’t be solved with better proximity tracking, critics say. No matter how effective these apps prove to be, they are just one cog in a pandemic response that also requires widely available testing, support for quarantine and isolation, and broadly observed social distancing measures.

The U.S. still doesn’t have nearly enough coronavirus tests to go around, which makes contact tracing far less effective. A Harvard research group estimates the U.S. will need to test about 5 million people a day (PDF) this summer and later ramp that number up to 20 million a day; to date, only about 12.6 million people have been tested over the course of several months.

“Rather than pinning our hopes on an inferior tech solution, I’d prefer to see us invest public resources in deploying testing and employing people to perform manual contact tracing that can generate more reliable data about who is likely to have been exposed to the virus,” says Justin Brookman, CR’s director of consumer privacy and technology.

“We’re capable of doing contact tracing without any of this technology,” says Marcus Plesica, chief medical officer for the Association of State and Territorial Health Officials. “We just do it in the old fashioned way we’ve done for decades.”