Seite - 272 - in VULNERABLE - The Law, Policy and Ethics of COVID-19

VULNERABLE272 for manual contact tracing even with lower uptake. Where a separate function of a contact-tracing app is to collect self-reported symptom data for AI, a much smaller percentage of users may still make the data collected useful for research purposes, even if that small percent- age renders the contact-tracing functions ineffective. Most governments that have adopted contact-tracing apps have made them voluntary, in spite of the need for significant uptake to be effective. Requiring all citizens to download and use an app that col- lects information about their movements or proximity to other indi- viduals raises serious civil liberties concerns. It is also not practicable or equitable in a context in which not every member of the population has a cellphone. While data from Statistics Canada suggest that a high number of Canadians have smartphones, the number of smartphones is significantly lower for seniors, with only 60.4% of those over the age of 65 reporting having a smartphone.20 Beyond this, the limita- tions of relying on cellphone-collected data to guide policy have been revealed in early experiments that used such data to guide local polic- ing and other municipal services, for example.21 Privacy It is evident that the architecture of contact-tracing apps has been influenced by privacy concerns. Initial use of GPS data raised the potential for surveillance. Bluetooth-based solutions aim to solve the surveillance problem by not capturing or sharing location data. Instead, when two phones using the Bluetooth-enabled app are in proximity, the devices exchange digital “tokens” that are stored only on users’ phones. These tokens record data about the device encoun- tered, as well as the degree of proximity and the duration of the encounter. A concern with Bluetooth models is that even de-identi- fied contact data could be used to create a “social graph” mapping an 20. “Smartphone Use and Smartphone Habits by Gender and Age Group” (last mod- ified 27 May 2020), online: Statistics Canada <doi.org/10.25318/2210011501-eng>. 21. See for example Kate Crawford, “The Hidden Biases in Big Data” (1 April 2013), online: Harvard Business Review <hbr.org/2013/04/the-hidden-biases-in-big- data>; Stephen Goldsmith & Susan Crawford, “The Responsive City: Engaging Communities through Data-smart Governance” (2014) 39:3 J Urban Affairs 458; Alred Tat-Kei Ho et al, “Big Data and Local Performance Management: The Experience of Kansas City, Missouri” In Yu-Che Chen & Michael Ahn, eds, Routledge  Handbook  on  Information  Technology  in  Government (Routledge, 2017) 95.