Despite privacy concerns, and even as the jury is out on the effectiveness of digital contact tracing, the federal government and several provinces are investigating citizen-tracking cellphone apps on the chance they will help reduce the spread of COVID-19 and lead to the reopening of the economy.
“There's some serious questions about whether or not it would be as effective as people are suggesting,” said Jason Millar, a University of Ottawa professor who holds the Canada Research Chair in the Ethical Engineering of Robotics and Artificial Intelligence. “It's possible (the apps) would be fantastic at contact tracing, but it's just not clear; we don't have that data. So right now, it's kind of a hope.”
Digital contact tracing apps track the movements of people using the embedded geolocation and Bluetooth systems. Simplistically, each phone emits unique pings or identifiers. When pings detect each other, they use Bluetooth signal strength to estimate whether the phones’ users came close enough, for long enough, to potentially pass on the virus.
Later, if either person becomes infected, the app uses geolocation data (stored on the phone or on an external server) to trace everywhere that person went prior to the onset of symptoms. The system then alerts everyone who pinged within range of an infected person during the contagious period that they have been potentially exposed.
Inherent to its success is an assumption geolocation and Bluetooth distance estimates are accurate. Relying on the phone’s built-in GPS and cell tower triangulation, geolocation can be incredibly accurate outdoors, except when it’s not. Indoors, all bets are off. The Bluetooth signal detection is also imperfect. Signal strength is affected by many things besides distance, including cellphone quality, solid objects between users, and interference from other Bluetooth devices.
So, some technical challenges to start.
Nevermind that, in Canada, people have been doing a pretty good job at social distancing without a tracing app. People are staying home and self-isolating; our hospital systems have not been overwhelmed; collectively, we’re flattening the proverbial infection curve. Ask any designer, introducing a new element into a working system could potentially make things worse rather than better.
“Throwing an app into the mix is not just an additional positive, risk-free layer of vigilance,” said Millar. “Because I think that's how technology is often portrayed.”
At this stage of the pandemic, the premise of keeping tabs on people who have tested positive as a measure to safeguard the rest is flawed.
While the app may help health-care workers track backwards through an infected person’s contacts to pre-symptom onset – digital tracking being more reliable than human memory – it could also overload them with “close contacts” misidentified by the imprecision of the technology. App users might then be flooded with anxiety-causing notifications (Alert: You may have been exposed to the virus – self-isolate.) when the infected person they were potentially exposed to three days ago was a passenger in the car next to them at a stoplight.
Add to this, we don’t know the true infection rate. Asymptomatic individuals can self-isolate without being diagnosed. A Stanford University study in Santa Clara, Calif., found up to 85 times more people were infected than had tested positive for COVID-19. When widespread community infection is coupled with limited testing capacity, as in many parts of Canada, knowing who and where the lab-tested positive people are delivers a false sense of security. “(A) proximity tracing system just does not work in that context,” said Don Willison, a University of Toronto professor at the Institute of Health Policy, Management and Evaluation.
If governments in Canada rush to use digital contact tracing as a pandemic-fighting measure by, say, mandating their public health departments to implement it, the move will be without consensus from scientists and medical experts that the apps will work as intended.
In fact, a review by the artificial intelligence, data and society ethics research watchdogs at the Ada Lovelace Institute in the U.K. last week panned the concept: “… mandating use of a digital contact tracing app is unlikely to be effective, enforceable or enjoy public support” and the “significant technical limitations, and deep social risks” far outweigh any value.
“We won't have time for widespread public consultation,” said Willison. People need to be able to express their concerns so the government can mitigate them in a transparent, trustworthy process of “focused engagement.”
Privacy is a primary concern for most people. No digital tracing app can guarantee complete anonymity.
“We have a strong privacy culture in Canada,” said Sean Holman, a journalism professor and freedom of information researcher at Mount Royal University in Calgary. “If people don't know something about you, they can't control you. If they can't control you, then you can feel more certain about your own actions, within your own sphere.”
If Canada, like some European nations, follows the lead of early adopters China, Taiwan, South Korea and Israel, which have forced myriad digital-tracker and tracing-type measures on their populations, it will have to sculpt a more privacy-sensitive iteration to get buy-in from the public.
“We are looking at what could be helpful for us, mindful of the privacy aspects we're all concerned about,” Dr. Bonnie Henry, provincial health officer for B.C., said last week. B.C. and Alberta are studying digital tracing options, including Singapore’s TraceTogether, which was a veritable flop with only 13 per cent of Singaporeans choosing to use it.
Willison said estimates range on the minimum participation rate required to be effective as being anywhere between 40 per cent and 75 per cent. “I’m more inclined to believe that 75 per cent,” Willison said. “If you have a lot of people who've chosen not to participate because they don't trust the system, it isn't going to work, you're going to have too many false negatives.”
The federal government has announced it is exploring options, including made-in-Canada versions. Most companies developing digital contact tracing use Bluetooth proximity tracing, which would, among other things, monitor people with the disease and identify high-risk COVID-19 hot spots so healthy people can avoid them on their way to and from work.
Is this wise? Pinpointing high-risk people and hot-spot locations could stoke a dangerous voyeuristic hobby (Track the infection in your community!) or even fuel vigilantism (the healthy versus the sick).
Under South Korea’s mandatory tracing system, the identity of an infected person is shared with government, posted publicly, and sent to every person they came in contact with over their infection period. Extreme cases of harassment caused the National Human Rights Commission of Korea to caution authorities to stop oversharing people’s details because they had become “the target of criticism, taunts and hatred online.”
Canada’s privacy commissioner Daniel Therrien recently stipulated every “privacy-impactful” measure during this pandemic must be more than useful, it must be “evidence-based,” “likely to be effective,” and mindful of “disproportionate impacts on vulnerable populations.”
As various versions stand proposed, digital tracing apps don’t check those boxes.
“It's not uncommon to point to technology and say, ‘Hey, technology can fix this.’ Technology is magical in a way,” said Millar. “It's almost like we imagine we can overcome problems that are hard to solve otherwise.”
These are extraordinary times. People are scared and want solutions.
“We're willing to do an awful lot to resolve (the) feeling of uncertainty, and uncontrollability,” said Holman. “Are people going to be willing to give up the control and certainty that privacy provides in favour of the control and certainty that surveillance will provide in a public health crisis?”
Willison’s research in non-crisis times revealed people were most willing to allow use of their personal health data for health research or quality improvement and that doctors, hospitals, and university researchers were perceived as the most trustworthy data keepers. “It was a matter (of) who uses it and how,” said Willison.
Which leads back to the privacy commissioner’s framework report that all data collected should be direct-to-purpose and destroyed after the crisis is over, meaning, no third-party use, no massive database of personal health information stored on private sector servers awaiting future commercialization opportunities.
“In the context of your exposure to the virus, what would be the potential social consequences of that information being used by parties to socially sort you? Asked Willison. “Say you're (deemed) not eligible for insurance, and not eligible for employment, or whatever it may be. That's the worry.”
Whatever Canadians decide, governments must tread with care. Any action perceived as heavy-handed or overreaching could meet a swift and punishing backlash.
“People tolerate measures that they wouldn't normally tolerate, and they do that upon a trust basis,” said Mark Warren, University of British Columbia political scientist and holder of the Merilees Chair for the Study of Democracy. “So governments gain more power because of the issue, but they need to use that power very carefully so they don't damage the public trust that they hold.”
There is a pervasive and unrealistic optimism many people have about technologic solutions, as if all the troublesome variables can be anticipated and programmed into submission in a lab. “That's not how technology works,” concludes the University of Ottawa’s Jason Millar. “We know when you put technology into society, things get messy.”
Comments